Can soil and tree carbon sequestration maintain zero net emissions grazing?

Farm summary

Jigsaw Farms is a prime lamb, premium wool, and cattle enterprise located in south-west Victoria. The farm was comprised of two properties in 2021: Hensley Park (−37°62′S, 142°01′E) and Melville Forest (−37°52′S, 141°82′E), 2133.1 ha and 1284.4 ha in area, respectively. At the time of purchase by the current owners (Mark Wootton and Eve Kantor) in the late 1990s, the stocking rate was five dry sheep equivalents (DSE) per hectare. Following pasture and soil improvement, this has increased to 25 DSE in some of the highest stocked paddocks. Average annual rainfall is 617.5 mm, predominantly in winter and spring. Rainfall variability in this region is relatively high. Rainfall was 559.5 mm in 2021.

By 1880, most native vegetation in this region had been cleared for agricultural production, with larger trees left in some areas for shade for animals. Starting in 1987, previous and current owners of Jigsaw Farms identified the need to restore biodiversity, manage salinity and provide shade and shelter for livestock, through the re-establishment of trees across the landscape. By 2021, a total of 16.9% of Hensley Park and 15.8% of Melville Forest had been established to environmental plantings or agroforestry. Since the publication of the original assessment (Doran-Browne et al. 2018), the owners have reduced the area under their management from 4900 ha to 3417.5 ha, with an associated reduction in total sheep and beef numbers, and 52.7 ha of trees have been planted. Changes have also occurred in parameters to assess emissions and sequestration, for example the GWP of methane has increased from 21 to 28 (IPCC 2014; DCCEW 2024) and the the Full Carbon Accounting Model (FullCAM) to estimate sequestration has been updated.

Tree carbon sequestration

Planted trees include agroforestry plantings of species not native to the area and intended for future harvest, as well as permanent revegetation (environmental plantings – a mix of native Eucalyptus and Acacia species). Agroforestry plantings were mostly planted in blocks between 2002 and 2010 and are predominantly spotted gum (Corymbia maculata, 72%) and Sydney blue gum (E. saligna, 19%) and a small area of mahogany gum (E. botryoides), red ironbark (E. tricarpa) and river red gum (E. camaldulensis), using tubestock seedlings at a density of 1000 stems per ha. These plantings were thinned mechanically, leaving about 500 stems per ha, with thinned trees left on site. Areas of permanent revegetation are a mix of trees and shrubs, using tubestock seedling or direct seeding. The previous owners of Melville Forest participated in the Potter Farmland Plan¹ and established 33 ha of permanent revegetation during the period 1987–1995. Since Jigsaw Farms acquired the property, the area of permanent revegetation more than trebled, to a total of 112 ha as of 2022 – some planted in blocks and others in belts. Prior to 1997, the proportions of trees and shrubs were 70:30; since then, the ratio has been 60:40.

The FullCAM model (2020 Public Release version) (Richards and Evans 2004) was used to estimate carbon sequestration in trees and shrubs. FullCAM is a dynamic model that incorporates spatial variation in site conditions and daily or monthly variation in weather conditions to model vegetation growth. It has calibrations to estimate carbon sequestration for different species, planting densities and planting configurations (blocks or belts). Modelling of carbon stocks in woody vegetation using FullCAM is carried out for carbon estimation areas (CEAs). These are areas of at least 0.2 ha with uniform site characteristics, planted at the same time with the same mix of plant species, and established and managed under the same regime, with ‘forest potential’ (the potential to reach two metres or more in height and crown cover of at least 20% of the land). Analysis with Google Earth and field observations indicated that the total stocked area of agroforestry and permanent revegetation of 562 ha (plantings that had achieved or could achieve forest cover of > 20% canopy cover) was less than the mapped area of plantings of 612 ha. The difference was due to dams located within mapped areas of tree plantings, some waterways mapped as tree planting that were fenced and protected but only partially revegetated, and areas of tree planting that either failed or were unlikely to achieve forest cover. At Hensley Park, the tree plantings were mapped into 99 CEAs ranging in size from 0.2 ha to 25.0 ha, with a total area of 360.4 ha. At Melville Forest, there were 56 CEAs ranging in size from 0.5 ha to 21.0 ha, with a total area of 202.4 ha. All CEAs were modelled with a start date of July 1st in the year the trees were established. The model run for each CEA started from the planting date until 2046, using a modelling point that was in the approximate centre of the CEA.

Two different approaches were used to assess carbon sequestration in trees: modelling with FullCAM and modelling adjusted with field measurements. FullCAM is a national model based on measurements of carbon stocks in trees of different ages, adjusted for climatic and soil factors and the estimated maximum forest biomass for any location in Australia. Its calibrations provide an average estimate of carbon sequestration over a wide range of sites. Sequestration rates at a given site may vary from this average depending on vegetation management and local site conditions. To assess the local accuracy of FullCAM output, field estimates were generated using tree and shrub diameter and height data collected in plots of varying size across both properties. Thirteen sites were assessed in the agroforestry plantings in August 2022 and 11 sites in permanent revegetation areas in April 2023. In agroforestry plantings, CEAs were 17–20 years of age. Measurements involved point sampling of basal area of the stands, area-based plots to measure stocking, and sampling of stand heights. Plots 20 m × 20 m were established at random in each CEA. Basal area was measured at the centre of the plot using a prism (2-factor). The number of live trees in the plot was recorded, and stand height was measured. Permanent revegetation was largely in belts and strips and assessed in randomly located belt transects 5 m wide and of variable length (11 m to 50 m), depending on the size and shape of the planting. At each plot, the stem diameter of all standing live trees that were at least 5 cm in diameter was measured and the species group recorded. Transect area was 1.3% of the area of the plantings. Data from both planting types were converted to estimates of carbon per unit area. For permanent revegetation above-ground, live biomass for each plot was estimated from stem diameter measurements using allometric models for each species group (Paul et al. 2016).

For the agroforestry plantings, field data was used to recalibrate the FullCAM Tree Yield Formula by adjusting the M value (the maximum above-ground biomass (t dry matter/ha) in undisturbed native vegetation) and parameter r (the site-productivity-dependent, non-endemic species multiplier) for each CEA to scale the modelled above-ground biomass to match the measured data. For another 18 CEAs, basal area and stocking were inferred from the measured CEAs considering their age, management history and geographic relationship and adjusted the parameter r in FullCAM. For the remaining 14 CEAs a user-defined calibration was used based on sugar gum (E. cladocalyx), but with the Tree Yield Formula (TYF) in FullCAM altered, using published parameters for ‘OtherEuc’, that predominantly comprises the mix of species E. cladocalyx, E. camaldulensis, E. saligna/botryoides, Corymbia maculata, and E. pilularis.

Soil carbon sequestration

Soil carbon data was sourced from farm records. Selected paddocks were first sampled at 0–10 cm depth between 2008 and 2018, and again in 2021 (Fig. 1), but were not included in the balance of GHGe. Soil data provided by Jigsaw Farms showed increases in soil carbon (t C/ha) in some paddocks (Carter, Quarry, and Stone) at 0–10 cm. However, statistical analysis suggested there was no obvious trend in soil carbon data (P value = 0.42). Henceforth we could not include annual soil carbon sequestration in the 2021 assessment, as it could not be concluded from the available evidence that changes in soil carbon were due to farm management activities or natural fluctuations in response to climate variability.

Fig. 1.

Soil carbon (t carbon/ha) and standard deviation from selected paddocks on Jigsaw Farms between 2008 and 2018, and in 2021, sampled at 0–10 cm.

Greenhouse gas emissions

Sheep and cattle numbers were sourced from farm records and modelled using the GrassGro model (Clark et al. 2000; Moore and Herrmann 2023) and reconciled with farm sales and purchase data. Climate data used in the GrassGro model was sourced from the SILO database (Jeffrey et al. 2001), based on Hamilton Airport (−37°39′S, 142°04′E). Data on fuel use, electricity, fertiliser, herbicides and purchased stock and feed were sourced from farm records. The emissions from the enterprise and the emissions intensity of beef, sheep meat and wool were calculated using the sheep and beef greenhouse accounting framework tool (SB-GAFv2.3) (Doran-Browne et al. 2018). The SB-GAF tool calculates Scope one and two emissions using the IPCC-approved Australian National Greenhouse Gas Inventory (NGGI) method (DCCEW 2024), whereas pre-farm Scope three emissions align with emission factors in the Ecoinvent database (Wernet et al. 2016). The tool calculates the farm greenhouse gas account within a pre-farm to farm gate life cycle assesment (LCA) boundary, as per ISO 14040 and ISO 14067 standards. Pre-farm emissions from fuel, electricity, feed, fertiliser, herbicides and pesticides were allocated equally between enterprises based on DSE. All emissions were converted to CO₂-e using the GWP of 28 for methane (CH₄) and 265 for nitrous oxide (N₂O), as per the NGGI (DCCEW 2024). Scope three emissions were included in the total greenhouse gas balance for the property and the calculation of emissions for agricultural products, in accordance with updated GHGe accounting protocols (DCCEW 2024). Allocation between meat and wool for the sheep enterprise was based on the protein allocation method of (Wiedemann et al. 2015, 2016b). The net greenhouse gas balance for Jigsaw farms was calculated for the 2021 calendar year, as the sum of all Scope one, Scope two and Scope three emissions, minus the annual change in tree carbon sequestered from agroforestry and permanent revegetation calculated for 2021.

Carbon sequestration in trees

Models validated with on-ground measurements, determined that between 2006 and 2021, Jigsaw Farms sequestered a total of −128,032 t CO₂-e, distributed as −75,170 t CO₂-e and −52,830 t CO₂-e at Hensley Park and Melville Forest, respectively; and −83,490 t CO₂-e and −44,510 t CO₂-e in the agroforestry and permanent revegetation plantings, respectively (Fig. 2). Annual sequestration rates peaked at −11,434 t CO₂-e per year in 2011 and declined thereafter as trees matured (Figs 2, 3). Modelled sequestration varied annually, largely driven by climate variability. Average annual sequestration per hectare was more than double in agroforestry plantings than permanent revegetation due to the latter having a higher proportion of shrubs with multiple stems and low maximum height in revegetation areas. During the 2021 calendar year, trees sequestered a total of −6704 t CO₂-e (Tables 1, 2), with a 10-year running mean of sequestration (2012–2021) of −8024 t CO₂-e (Fig. 3).

Fig. 2.

Modelled annual change in tree carbon sequestration (t CO₂-e/year) on Jigsaw Farms cumulative carbon stocks (t CO₂-e) from sequestration from 2006 to 2021 from agroforestry and permanent revegetation across Hensley Park and Melville Forest.

Fig. 3.

Modelled annual change in tree carbon sequestration (t CO₂-e/year) from agroforestry and permanent revegetation plantings on Jigsaw Farms from 2011 to 2021 and the CO₂, CH₄, N₂O and total gross GHGe produced by Jigsaw Farms in 2021.

Comparing the sequestration results and carbon balance of this study with those of the previous study on the Hensley Park property (Doran-Browne et al. 2018). This study indicated a lower planted area (360 vs 380 ha) because the areas were adjusted for actual stocked area, but higher sequestration rates. This resulted in similar total amounts of carbon sequestered on this property between 2006 and 2014 (−35,480 t CO₂-e for Doran-Browne et al. and −38,860 t CO₂-e in our study). Both studies found that the peak rate of carbon sequestration by planted trees occurred in 2011 (Table 1).

Emissions from livestock, farm operations and inputs

It was estimated that during the 2021 calendar year, emissions from Jigsaw farms operations totalled 10,870 t CO₂-e, with 9830 t CO₂-e directly on-farm (Scope one and two) and 1040 t CO₂-e in pre-farm inputs (Scope three, see Table 3). On-farm emissions were predominantly enteric CH₄ (78.2% of Scope one emissions), followed by lime (7.3%), N₂O from animal excreta (4.7%), CH₄ from manure management (4.1%), N₂O from leaching (3.4%) and CO₂ from fuel use (1.2%). Scope two emissions from the use of electricity totalled 64 t CO₂-e (Table 3). This was relatively low because Jigsaw Farms has a solar power installation, reducing its reliance on grid electricity. Scope three emissions were included in the calculation of intensity of products, and property emissions. The largest sources of Scope three emissions were purchased feed (72.7%), fertiliser (13.8%), livestock purchase (4.8%) and lime (4.3%) (Table 3).

During 2021, Jigsaw Farms sold 165,805 t beef, 593,291 t sheep meat and 126,531 t greasy wool (Table 2). The emissions intensity (excluding sequestration) of each product was calculated to be 11.6 kg CO₂-e/kg liveweight for beef, 8.5 kg CO₂-e/kg liveweight for sheep and 30.8 kg CO₂-e/kg of greasy wool (Table 2).

Emissions balance

In 2021, total gross emissions were 10,870 t CO₂-e, minus total sequestration of −6704 t CO₂-e, resulting in net emissions of 4166 t CO₂-e (Table 3). Tree sequestration therefore offset 61.7% of total emissions (Table 1; Fig. 3). Allocating sequestration offsets equally between the products reduced emissions intensity to 3.5 kg CO₂-e/kg liveweight of beef, 3.4 kg CO₂-e/kg liveweight of sheep and 12.3 kg CO₂-e/kg greasy wool, respectively (Table 2).

If emissions in the early 2010s were similar to those estimated for 2021, the property would have been carbon neutral for a period of about 5 years from 2008 to 2012. The previous carbon assessment (Doran-Browne et al. 2018) for the property was for the Hensley park property only. A comparison indicates that on average, livestock emissions and sequestration were respectively 2.6 and 1.6 times lower in Doran-Browne et al. (2018), than calculated in the current assessment (Table 1). The ratio of tree carbon sequestration to livestock emissions (calculated as % carbon inset) was clearly greater for the Hensley Park property (21% trees, 2010–2014) than Melville Forest (15.8% trees in 2021). The study by Doran-Browne et al. (2018), and sequestration data from this study estimate that Jigsaw Farms was net zero between 2010 and 2017 (shaded in Table 1).