Integrating dual-purpose crops mitigates feedbase risk and facilitates improved lamb production systems across environments: a whole-farm modelling analysis

Keywords: APSIM, canola, feedbase, lambing time, supplementation, wheat, whole-farm systems, winter feed gap.

Many livestock grazing systems are impacted by periods of insufficient on-farm feed supply that either requires expensive supplementary feed or reduces the safe carrying capacity of livestock at certain times of the year, which limits total farm productivity potential (Moore et al. 2009; Bell et al. 2018). A changing climate has also made the frequency and severity of on-farm feed shortages more common and unpredictable (Bell et al. 2018; Henry et al. 2018; Godde et al. 2021); hence, ruminant livestock systems will need to adapt to reduce the risk of farm income losses in the face of increasing climate variability (Bell et al. 2014). Feed gaps, where livestock feed demand exceeds supply, predominate in the winter and summer/early autumn in many Australian livestock systems. These feed gaps not only lower production potential and increase input costs, but also reduce livestock efficiency, as animals will often lose and then regain weight, which requires ~50% more energy than an animal that gains and maintains the same weight (Moore et al. 2009). Designing forage systems that reduce the frequency and intensity of feed gaps can bring about significant gains in productivity and efficiency of livestock systems, and reduce risks of environmental degradation.

These feedbase limitations also affect the timing of livestock management activities, particularly the timing of reproduction and sale of livestock, which are timed to avoid periods more prone to feed gaps. In southern Australian sheep systems, lambs are typically born in late winter or early spring to avoid risks of winter feed gaps and to better utilise the more reliable spring pasture supply (Freer et al. 1994; Croker et al. 2009). Although this system has been adopted to best match the livestock feed demand with pasture forage supply at lambing time, it does introduce some limitations and constraints to livestock production systems. For example, later spring-lambing systems allow less opportunity for lambs to meet target liveweight (LW) conditions before late spring/early summer, when feed nutritive value is in decline and expensive supplementary feed is required to maintain consistent growth (Reeve and Sharkey 1980; Freer et al. 1994). Lambing at this time may also predispose lambs to a higher risk of internal parasites, notably Haemonchus contortus (also known as Barber’s pole worm) and flystrike (cutaneous myiasis) caused by the blowfly, Lucilia cuprina (Horton and Champion 2001), after weaning, especially in areas that receive a higher proportion of rainfall in the summer months, with elevated humidity favouring the lifecycle of these parasites (e.g. central, and northern New South Wales; NSW).

Dual-purpose crops provide an additional forage option that can help fill winter feed gaps and are becoming more commonly adopted in mixed crop–livestock systems in southern Australia (Moore 2009). Dual-purpose crops, such as wheat and canola, provide forage of high nutritive value during early autumn to mid-winter (May to August), which can provide several whole-farm benefits, including the spelling of slow winter-growing pastures (Virgona et al. 2008; Dove et al. 2015) and additional income from grain production, provided grazing is managed appropriately (Kirkegaard et al. 2008; Bell et al. 2015a). Previous analyses have shown that the integration of these dual-purpose crops to the feedbase can enable livestock producers to safely increase or maintain stocking rates over winter, without increasing supplementary feed requirements (McGrath et al. 2014). Compared with pasture-only livestock systems, allocating 10–20% of the farm area to dual-purpose wheat and canola can increase total farm sheep grazing days by 10–15% and farm profits by more than AUD$150/farm ha (Bell et al. 2015a). Some producers who have adopted dual-purpose crops have also begun to utilise the more reliable winter feed supply to shift lambing time earlier to the late autumn/winter period (McGrath et al. 2013). Simulation modelling in a uniform rainfall environment at Wagga Wagga, New South Wales, Australia, found that providing additional grazing area via access to dual-purpose wheat supported earlier lambing in autumn and increased lamb production, and decreased supplementary feed costs, resulting in higher livestock gross margins than a pasture-only system (McGrath et al. 2014). This suggests that the additional forage provided by dual-purpose crops could offer further systems benefits, and allow shifts in lambing time and an increase in stocking density, as winter feed gaps are filled. However, the geographical and environmental scope of previous modelling analyses has been limited, and more detailed analyses in the context of a whole-farm system are yet to be explored. There are also likely to be other potential legacy effects of the feedbase on livestock reproduction, such as body condition score (BCS) of ewes at joining and subsequent reproduction rates, and potential for improved animal health and welfare, which have also not been examined. Further evidence and understanding of the multiple benefits of dual-purpose crop integration to livestock systems and their interactions with livestock production management is required across a broader range of environments.

To address this, we used a simulation model of a whole-farm livestock enterprise that incorporates dual-purpose crops compared with those with the more common pasture-only feedbase to explore the potential impacts of dual-purpose crop integration in combination with spring- or autumn-lambing on: (1) livestock production, (2) the frequency of feed gaps, hence, supplementary feed requirements, and (3) the economic performance of a sheep-enterprise across four different environments in NSW, Australia. We modelled these systems across different stocking densities (based on pasture utilisation rates) to explore how these interacted with the different feedbase systems.

Twelve whole-farm model scenarios were designed and simulated across four locations spanning different agro-environments in NSW, Australia. These locations were chosen to capture a range of climatic conditions differing in total rainfall and rainfall distribution (summer rainfall %) including Armidale (high rainfall, summer-dominant), Gulargambone (medium rainfall, summer-dominant), Goulburn (high rainfall, uniform) and Temora (medium rainfall, uniform; Fig. 1). The 12 simulations involved a factorial of two lambing systems (autumn vs spring) by two feedbase types (100% pasture vs 75% pasture and 25% dual-purpose crops) by three stocking densities. These stocking densities were derived to correspond to feedbase utilisation rates of 30% (low), 40% (medium) and 50% (high) by computing the average proportion of forage growth that was consumed by livestock in the baseline pasture-only system at each location. The stocking densities at these locations varied due to their inherent differences in productivity of the pasture feedbase (see Supplementary Table S1).

Fig. 1.  Map of the four locations used in the whole-farm model scenarios in APSIM, including the average monthly rainfall (bars), and minimum (dotted line) and maximum (solid line) temperature.

Whole-farm systems simulations were developed using the Agricultural Production Systems Simulator version 7.10 (APSIM; Holzworth et al. 2014) via the shared binary protocol (CSIRO Common Modelling Protocol) of APSIM and GRAZPLAN modules (Moore et al. 2007). This allows for simulations that combine forage or dual-purpose crops, and pasture growth modules with modules that drive livestock grazing, growth and reproduction. The APSIM modelling platform was used instead of AusFarm, which has been used by others to conduct similar analyses (e.g. McGrath et al. 2014), to reduce some of the complexity involved to develop management rules for the whole-farm systems, as AusFarm uses its own unique modelling language. Nonetheless, the simulations modelled here used the same widely tested and validated models describing livestock production and reproduction (Freer et al. 1997), growth and nutritive value of temperate perennial pastures (Moore et al. 1997), and growth and yield of wheat and canola crops (www.apsim.info).

Two paired 30-year simulations were run for each scenario at each location (i.e. 30 independent simulations: 15 cycles for each paired simulation) for the years 1987–2017 and 1988–2018. These paired simulations were offset by one year from each other to capture legacies from the first to the second years, and allowed for the examination of climatic variation effects over the period.

All simulations were designed for a 1200-ha farm that included eight paddocks, each 150-ha in size. In the pasture-only scenarios, all eight paddocks were allocated to established permanent monoculture pastures, whereas in the scenarios with dual-purpose crops, six paddocks were allocated to established permanent monoculture pastures, and the other two paddocks were sown to dual-purpose crops each year (one sown to wheat and the other canola). To capture variation in soil types, farms were separated into two or three different land management units (LMU): LMU 1 represented soils of high fertility suitable for improved pasture and/or cropping; LMU 2 represented soils of moderate fertility suitable to improved pasture; and LMU 3 represented soils of lower fertility suitable for lower productivity native or naturalised pastures (Table 1). At Armidale and Goulburn, three of the paddocks allocated a LMU 3 were set to an ‘undulating’ landscape (8.3° slope), which better represented the geography of these locations that were located close to the Great Dividing Range. All other paddocks at these two locations and at Gulargambone and Temora were set to a ‘gentle’ landscape (4.2° slope). The Soils and Landscapes Grid of Australia (Grundy et al. 2015) and the SoilMapp iPad app (CSIRO 2020) were used to classify representative soils based on type and depth for each location, which were then used to select analogous soils found in the region from the APSoil database (https://www.apsim.info/apsim-model/apsoil/; Table 1).

Table 1.  Land management unit (LMU) contribution to farm area, associated soil types and their plant-available water-holding capacity (PAWC) at each of the four simulated locations in NSW.

The pasture types used for each location were selected based on regionally dominant perennial pastures, which were derived via consultation with industry advisors and information in the literature (McDonald 1998, 1999, 2004). In the scenarios with pasture-only, permanent pastures comprised of 100% phalaris at Armidale; 62.5% lucerne sown in LMU 1 and 2, and 37.5% Austrostipa sown in LMU 3 at Gulargambone; 62.5% phalaris sown in LMU 1 and 2, and 37.5% Microlaena sown in LMU 3 at Goulburn; and 62.5% lucerne sown in LMU 1 and 2, and 37.5% phalaris sown in LMU 2 at Temora. In the scenarios with dual-purpose crops, crops displaced the most productive pastures, resulting in 12.5% each of wheat and canola sown in LMU1, with the rest of the farm sown to pasture. At Armidale, phalaris still made up the entire pasture base. However, at all other locations that included two pasture species, the displacement of the most productive pastures with crops resulted in an even split of these pasture species on the farm (37.5% each), with improved pastures of winter active lucerne and phalaris sown in LMU 2, and native/naturalised pasture species of Austrostipa spp. and Microlaena sown in LMU 3. A formal validation was not undertaken for the simulated average monthly pasture growth, although the pasture models used have been widely tested and validated, and outputs were cross-checked with literature relevant to each location for sensibility (McDonald 1998, 1999, 2004; www.evergraze.com.au).

The simulation of dual-purpose crops was implemented by selecting two different wheat and canola cultivars specific to each location to match the optimal sowing window that allowed for maximum grazing and grain yield potential (based on flowering dates; Bell et al. 2015b; Lilley et al. 2015; Flohr et al. 2018; Supplementary Table S1). Crop sowing was triggered by 25 mm of rain received over a minimum 3 days, and the cultivar sown was dictated by the timing of the rain event. If a rain event did not occur, a forced dry-sowing event of the earlier maturing cultivar was triggered on the last day of the sowing window. In higher-rainfall, cooler locations (Armidale and Goulburn), a slower developing winter wheat (cv. Revenue) and canola (cv. Taurus) could be sown in late summer/early autumn, but if these could not be sown within the optimal sowing window, a faster winter wheat (cv. Wedgetail) and canola (cv. CBI406) was sown in mid/late autumn. At the lower rainfall, warmer locations (Gulargambone and Temora), faster developing winter wheat (cv. Wedgetail) and canola (cv. CBI406) were sown in early autumn, but if these could not be sown within the optimal window, then a medium developing spring wheat (cv. Gregory) and canola (cv. 46Y78) were sown in mid/late autumn (Supplementary Table S1).

Wheat and canola crops were reset for plant-available water, surface organic matter, and soil nitrogen (N) at the end of December for Armidale, Goulburn and Temora, and at the end of November for Gulargambone. These dates were chosen because they were within 14–30 days post-harvest in all simulated years. Resetting plant-available water shortly after harvest enabled soil water to accumulate with rainfall, as would occur in a normal fallow system. The average surface residue was set each year based on 80% of the average crop residue for wheat and 70% of the average crop residue for canola. Soil nitrogen was reset with 50 kg/ha of soil nitrate-N distributed evenly throughout the soil profile after harvest. All crops were fertilised with 100 kg N/ha at sowing, with an additional 100 kg N/ha applied (as urea) when the crop reached growth stage 29–31 (floral initiation; APSIM stage 4.9–5.1).

The opportunity for grazing of the dual-purpose crops closed at floral initiation, as it is known that grazing before this time has little or no impact on subsequent crop yield, whereas later grazing can induce yield penalties (Harrison et al. 2011; Kirkegaard et al. 2012; Sprague et al. 2015). The date that grazing ended was set by simulating an ungrazed crop and determining the 75th percentile for the date when wheat and canola reached stem elongation. These dates were kept constant for each crop at all locations and simulations, and applied as a crop grazing rule. At Armidale and Gulargambone (the northern locations), end grazing dates for wheat and canola were 11–22 August and 30 June−31 August, respectively. At Goulburn and Temora (the southern locations), end grazing dates for wheat and canola were 25–31 August and 10–18 July, respectively (Supplementary Table S1). To simulate the grain yields of the dual-purpose crops, paired simulations of an ‘ungrazed’ wheat and canola crop was run alongside the ‘grazed’ simulations based on the assumption that best grazing management applied in our simulations would not have detrimental effects on grain yield. A similar approach has been employed by others (e.g. Bell et al. 2015b; Lilley et al. 2015), because the complex processes influencing subsequent regrowth and grain yield recovery of dual-purpose crops are currently not simulated with high confidence in APSIM.

Consistent management of the livestock enterprise was maintained across all simulated scenarios and locations to avoid dramatically changing livestock demand among them, and hence, introduce bias or complexities when making intercomparisons among them. To allow each scenario to start at a common point, and hence, isolate the legacy effect of the different feedbase on subsequent year livestock productivity, the livestock enterprise was ‘reset’ at the end of each 2-year simulation cycle by culling and purchasing a new cohort of ewes. The livestock enterprise was a prime lamb enterprise consisting of a single-aged cohort of medium-sized Merino ewes joined to a Suffolk ram. Ewes were purchased at the age of 3 years with a LW of 50 kg and a BCS of 3.0, and culled at the age of 5 years. This action of culling and replacement occurred after weaning, but before joining, in early November in autumn-lambing systems, and mid-February in spring-lambing systems. In autumn-lambing systems, ewes were joined in mid-November and lambs were born towards the end of April. In spring-lambing systems, ewes were joined at the beginning of March and lambs were born in mid-August. In both systems, lambs were weaned at 98 days old and sold once they reached a target LW of 48 kg. Lambs that did not meet the target LW by 5 months post-weaning were sold at this time. In each lambing system, all other management activities (e.g. shearing and crutching) occurred at the same time relative to joining. More specific details of the livestock management activities are outlined in Supplementary Table S2.

In the simulations, the movement of animals between the pasture paddocks was predicated on available biomass (i.e. above-ground biomass not including surface litter or residues). When the available biomass in the ‘current paddock’ was reduced to ≤1 t DM/ha, animals were moved to the pasture paddock that contained the most biomass. After weaning, ewes and lambs were rotated around the farm in separate flocks using these same rules until all lambs were sold, at which point all paddocks became available again to the ewes. In the pasture-only scenarios, ewes had access to six pastures paddocks (LMU 2 and 3), and the lambs had access to the remaining paddocks (LMU 1). In the scenarios with dual-purpose crops, ewes had access to four pasture paddocks (LMU 2 and 3) plus the wheat stubble post-harvest (LMU 1), and the lambs had access to the remaining two pasture paddocks (LMU 2).

In the scenarios with dual-purpose crops, animals were permitted to graze the crops over the winter based on a series of crop grazing rules. Wheat grazing was allowed when crop biomass reached 1.5 t DM/ha, and canola grazing was allowed when the crop biomass reached 2.0 t DM/ha. The grazing of both crops ceased when biomass was reduced to ≤0.8 t DM/ha or when the crop reached the end grazing date (as previously described, to align with just prior to bud elongation). Depending on sowing date and season, crops could be grazed multiple times over the winter period, provided all conditions were met. After the wheat crop was harvested, ewes could graze wheat stubble for a maximum of 14 days, where they had access to grain supplement equivalent to 3% of average crop grain yield to approximate spilt grain. Canola stubble was not utilised by livestock.

Supplementary feeding was initiated in simulations to compute the unmet livestock demand for the different feedbase systems. Maintenance feeding of ewes was initiated when the average BCS of the flock was ≤2.5 or when pasture biomass in the ‘best paddock’ fell below 1.5 t DM/ha. Lambs were only supplemented after weaning when the average daily weight gain of the lamb cohort was ≤0.05 kg/day. Supplement demand (calculated as metabolisable energy (ME) demand) of the livestock enterprise was calculated based on the maintenance feeding of the wheat supplement provided (Eqn 1). The ME value, 13.8 MJ ME/kg DM, was based on the ME content of the wheat grain provided in the simulations using values from standard feed test reports (Agriculture Victoria 2021).

Wheat grain supplement was offered at a maximum rate of 1 kg/head.day to ewes and lambs when required; however, animals only consumed supplement at a rate needed to meet maintenance requirements, which was dependent on pasture availability and nutritive value. Supplement was not permitted during the crop grazing period in the scenarios with dual-purpose crops.

The gross margin of each of the scenarios was calculated on a fiscal year basis (July to June) using livestock sales and grain sales, and the associated input costs of these. Although there may be some small differences between locations in input costs (e.g. fertiliser inputs, pesticide applications) and/or commodity prices, to maintain consistency and the opportunity for intercomparison between locations, the costs and prices assumed remained constant.

The sale of lambs was calculated using market specifications for carcass weight (CWT) (18–22 kg CWT and 12–18 kg CWT), whereas the sale of cull ewes was based on a set price/kg CWT. A set transport cost of AUD$4/head, skin price of AUD$5/head and dressing percent of 41% was applied to all livestock sales. Replacement ewes were purchased at a set price (AUD$79/ewe). The cost of shearing and crutching was based on the award rate prior to 2020, as specified by the Shearing Contractors’ Association of Australia (www.scaa.org.au). The cost and income from wool was based on breeding ewes only (as lambs were not shorn or crutched between weaning and sale), with wool sales calculated from wool yielded from shearing only; the model was not set up to calculate crutched wool, which typically generates low relative value. Prices assumed for lamb and sheep meat, live sheep (replacement ewes), and wool prices were based on the average price over the past 10 years (2010–2020) accessed from the Meat and Livestock Australia market information statistics database (http://statistics.mla.com.au/Report/List) and Australian Wool Innovation. Wheat supplement was set at AUD$255/tonne, and costs incurred were calculated on an ‘as consumed’ basis. A livestock variable cost was set at AUD$25/breeding ewe, which included the total input costs for animal health (drenches and vaccines) and the labour costs of these, plus general labour associated with the care and movement of livestock on farm. A pasture and livestock infrastructure cost of AUD$300/pasture ha was also applied, which accounted for input costs associated with the maintenance of fences and other livestock infrastructure, and fertiliser for pastures (Supplementary Table S3).

In the systems with dual-purpose crops, crops provided an additional income to the farm enterprise. The wheat and canola crops were assumed to require a set crop variable cost of AUD$640/crop ha and AUD$750/crop ha, respectively, which included the total input cost of seed, fertiliser, pesticide applications, and costs of sowing and harvest contractors based on estimates from industry advisors. Crop grain yields were based on model simulations of ‘ungrazed’ wheat and canola over the same simulated years, but were adjusted to account for grain spillage and moisture content at harvest, with a 1.02 multiplier applied for canola and a 1.09 multiplier applied for wheat. Wheat price was set at AUD$255/tonne and canola was set at AUD$520/tonne.

Shifting from spring- to autumn-lambing increased feed deficits, and hence, supplement demand during late autumn and winter. Whereas for spring-lambing systems, supplement demand was higher in spring, summer and autumn (Fig. 2). In autumn-lambing systems, dual-purpose crops significantly improved winter feed supply and reduced supplement demand compared to a pasture-only feedbase in all simulated locations (Table 2; Fig. 2). The integration of dual-purpose crops was most beneficial at the high, uniform rainfall location of Goulburn, as the supplement required to meet livestock demand was reduced by ~60% compared with a system relying on pasture-only.

Fig. 2.  Average monthly supplement demand (GJ ME/ha) at Armidale (a–d), Goulburn (e–h), Gulargambone (i–l) and Temora (m–p), NSW, based on the supplementary feed requirements of ewes (dark blue) and lambs (light blue) in spring- and autumn-lambing systems, with pasture-only or with the integration of dual-purpose crops (DPC). Results presented for the ‘medium’ stocking density (i.e. 40% feedbase utilisation rates).

Table 2.  Simulated supplement required to meet livestock demand, and the percentage of that supplement fed to lambs under a spring- or autumn-lambing system with pasture-only or with the integration of dual-purpose crops (DPC) stocked at ‘low’, ‘medium’ and ‘high’ stocking densities (i.e. 30%, 40% and 50% feedbase utilisation rates, respectively) at four locations in NSW.

The impact of dual-purpose crop integration on the frequency and size of feed deficits, and hence, the amount of supplement required, varied across the different locations and altered the potential benefit in shifting lambing time to autumn among the study locations. Autumn-lambing systems with dual-purpose crops were most favourable in the higher-rainfall, cooler environments of Armidale and Goulburn. At these two locations, the autumn-lambing system with dual-purpose crops had a less severe feed gap during winter (Fig. 2a–h), and the supplement required was on average approximately one-third of that for the other scenarios tested (Table 2). At Goulburn in particular, the frequency and amount of supplement demand for the autumn-lambing system with dual-purpose crops was considerably less across all seasonal conditions, and across all three stocking densities (Fig. 3b). At Armidale, with more summer dominant rainfall, the effect of dual-purpose crops was less, with a similar frequency in supplement demand between all scenarios tested; although the supplement demand of the autumn-lambing system with dual-purpose crops was still marginally lower, especially at high stocking density (Table 2; Fig. 3a).

Fig. 3.  Cumulative distribution of annual supplement required to meet livestock demand for spring- (dotted) and autumn- (solid) lambing systems with pasture-only (green) or with the integration of dual-purpose crops (DPC; gold), and stocked at ‘low’, ‘medium’ and ‘high’ stocking densities (i.e. 30%, 40% and 50% feedbase utilisation rates, respectively) at Armidale (a), Goulburn (b), Gulargambone (c) and Temora (d), NSW.

The positive impact of dual-purpose crop integration in both autumn- and spring-lambing systems was less consistent in the lower-rainfall, warmer environments of Gulargambone and Temora. At Temora, the frequency and amount of supplement demand at low stocking density was marginally less for spring-lambing systems with dual-purpose crops, but as stocking density increased to medium and high stocking densities, the benefit of dual-purpose crops was similar for both autumn- and spring-lambing systems (Table 2; Fig. 3d). At this location, the spring-lambing system with pasture-only had the highest supplement demand of all scenarios tested (Table 2; Fig. 3d), with the greatest demand during the spring (Fig. 2m–p), when ewe demand was peak due to lambing and lactation. Dual-purpose crops reduced the amount of supplement demand in spring-lambing systems during spring, and in autumn-lambing systems in late summer/early autumn, despite the crops not being available for grazing in these seasonal windows (Fig. 2m–p). The benefit of dual-purpose crops at Temora was most apparent at medium and high stocking densities (Fig. 3d).

At Gulargambone, the location with lower and summer-dominant rainfall, there were still considerable forage deficits from May to July, indicating that autumn-lambing systems with dual-purpose crops are less viable at this location, even though the crops improve winter feed supply (Fig. 2k–l). For the autumn-lambing system, the displacement of pasture with dual-purpose crops also induced more frequent feed deficits in summer (January and February). Integrating dual-purpose crops into spring-lambing systems may also be problematic at this location, as the displacement of pasture for crops also induced more frequent feed deficits in summer and early autumn; hence, there was insufficient feed to adequately support the demands of the growing lambs and dry ewes (Fig. 2i, j). As a result, spring-lambing systems with pasture-only more often had lower supplement demand and frequency of supplement demand across all seasonal conditions (Fig. 3c), and hence, were most favoured at this location, especially at low and medium stocking densities. However, at high stocking densities, the autumn-lambing system that integrated dual-purpose crops performed more favourably (9% less supplement demand) than one relying on pasture alone (Table 2). At the high stocking density, the dual-purpose crops mitigated some of the winter feed gap, whereas the spring-lambing system with pasture-only incurred a much larger feed deficit in the summer (Table 2).

The use of supplement among different livestock classes also shifted with different lambing systems. Because autumn-born lambs were sold in early summer, very little supplement was fed to lambs, and this was further reduced with the integration of dual-purpose crops (Table 2). In contrast, a high proportion of supplement was required by spring-born lambs over the late summer and early autumn (February to March), particularly at Goulburn and Armidale (Table 2; Fig. 2a–h). At these high rainfall sites, the percent of supplement fed to autumn-born lambs was very low (<6%) compared with spring-born lambs (>32%), because forage quantity and nutritive value was lower at these locations over summer (Table 2; Fig. 2a–h); hence, the supplement demand of spring-born lambs was higher over the summer months. The use of lucerne pasture that is summer-active, at the lower rainfall sites of Gulargambone and Temora, offset supplement demand over summer for spring-born lambs (Fig. 2i–p).

The stocking density employed to achieve different feedbase utilisation rates interacted strongly with the feedbase employed and the potential benefits of dual-purpose crop integration. The frequency of supplement demand of each system in the different seasonal conditions increased as stocking density increased (Fig. 3). At high stocking densities, integrating dual-purpose crops led to a lower relative increase in supplement demand compared with systems with a pasture-only feedbase. This was especially the case for autumn-lambing systems (Table 2), where greater feed demand of lambing/lactating ewes led to more frequent feed deficits under higher stocking densities, but integrating dual-purpose crops helped to mitigate these deficits in late autumn and winter (Fig. 3). At low stocking densities, differences in supplement demand were fewer between systems that integrated dual-purpose crops compared with the pasture-only systems, because there were fewer instances in the pasture-only systems where feed demand of livestock exceeded feed supply provided by pasture (Fig. 3).

Although dual-purpose crops helped fill the winter feed gap in autumn-lambing systems, at any given stocking density, this did not always result in higher livestock production metrics (lambs or ewes) compared with those relying on pasture-only. They also provided no additional production benefit to spring-lambing systems (Table 3). This was largely because animals were supplemented to meet maintenance requirements when forage was insufficient, which would have masked any potential impacts of feed availability from dual-purpose crops that may have otherwise occurred. Lamb production from spring-lambing systems that integrated dual-purpose crops was reduced in some cases compared with pasture-only systems (Table 3). This was because spring-born lambs did not graze the dual-purpose crops of high nutritive value during winter and only had access to lower productivity pastures from spring to early autumn, because the better pasture areas were displaced by crops (i.e. grown on the most productive soil types; LMU1).

Table 3.  Simulated lamb production outputs for spring- and autumn-lambing systems with pasture-only or with the integration of dual-purpose crops (DPC) stocked at ‘low’, ‘medium’ and ‘high’ stocking densities (30%, 40% and 50% feedbase utilisation rates, respectively) at four locations in NSW.

At all locations, regardless of feedbase, shifting from a spring- to an autumn-lambing system reduced lamb weaning rates on average by ~13% (data not shown). This reflects the seasonality of ewe ovulation and oestrus that peaks in autumn when joining is scheduled for spring-lambing. Autumn-born lambs were also ~5% lighter at weaning (data not shown). However, shifting from a spring- to an autumn-lambing system brought about increases in individual lamb growth and marketability. On average, autumn-born lambs reached target LW ~16 days sooner, 32% more lambs reached target LW and carcass weights were 9% higher than spring-born lambs, although the lower weaning rates meant total lamb production/ha was ~4% less (Table 3).

Shifting from a spring- to an autumn-lambing system that was made viable from integration of dual-purpose crops, also significantly improved ewe body condition at joining and reproduction in their second lambing cycle. On average, autumn-lambing ewes had improved body condition (by 0.8 BCS) after their first lambing, compared with spring-lambing ewes that more often had lower body condition (−0.1 BCS; Table 4). This difference in ewe condition translated into an improvement in lamb weaning rates in the second lambing cycle with weaning rates in autumn-lambing systems increasing on average by 7.5% compared with only 0.1% in spring-lambing systems. At some locations, the weaning rates in spring-lambing systems decreased by up to 6% in the second lambing cycle, indicating that these systems were more prone to declining reproductive performance over time (Table 4) due to a lower BCS of 2.5–3.0 for these ewes (data not shown).

Table 4.  Difference in ewe body condition and reproductive performance (based on lamb weaning rates) in their second lambing cycle compared with those in their first year (i.e. legacy effect) for spring- and autumn-lambing systems with pasture-only or with the integration of dual-purpose crops (DPC) stocked at ‘low’, ‘medium’ and ‘high’ stocking densities (30%, 40% and 50% feedbase utilisation rates, respectively) at four locations in NSW.

Livestock production varied across most locations for the lambing time and forage system combinations with feedbase deficits limiting individual lamb production, and ewe body condition and reproduction in the second lambing cycle at different times of the year. At the higher-rainfall, cooler environments of Armidale and Goulburn, autumn-lambing systems with pasture-only had the highest percent of lambs sold at target LW, and lambs were sold sooner, compared with the other scenarios tested, as lambs had access to the most productive pastures and required little or no supplementation. However, at higher stocking density, differences in lamb production metrics were marginal between the autumn-lambing systems (Table 3). Spring-born lambs experienced feed deficits in summer and early autumn, and required supplementation (Fig. 2), which limited their growth rates and final carcass weights compared with autumn-born lambs. As a result, total lamb production was marginally higher for autumn-lambing systems at medium and high stocking density (Table 3), despite higher weaning rates in spring-lambing systems (data not shown).

At the lower-rainfall, warmer environments of Gulargambone and Temora, integrating dual-purpose crops provided more lamb production benefits, particularly in autumn-lambing systems. At Gulargambone, at low and medium stocking density, integrating dual-purpose crops allowed autumn-born lambs to reach target LW sooner compared with all other scenarios, but at high stocking density there was no difference between the autumn-lambing systems (Table 3). In general, the percent of lambs that reached target LW, and carcass weights, were better for autumn-lambing systems compared with spring-lambing systems, but total lamb production was always lower (Table 3) due to the higher weaning rates in spring-lambing systems (data not shown). Differences in these lamb production metrics for the two lambing systems were more marginal at low stocking density, but as stocking density increased, the percent of lambs that reached target LW declined significantly in the spring-lambing systems (Table 3). At Temora, at low and medium stocking density, the number of days to reach target LW, percent of lambs that met target LW, and carcass weights were similar for the autumn-lambing systems and the spring-lambing system with pasture-only. However, at high stocking density, these lamb production metrics were marginally better for the autumn-lambing system with pasture-only (Table 3), as these lambs had access to the most productive pastures and did not experience feed deficits in summer. At this location, total lamb production was highest for the spring-lambing systems due to the higher lamb weaning rates (data not shown), as carcass weights were only marginally higher for autumn-lambing systems (Table 3). Integrating dual-purpose crops into spring-lambing systems benefitted total lamb production at low and medium stocking density, but at high stocking density the pasture-only system was better, as feed deficits in summer/early autumn in the dual-purpose crop system limited individual lamb production (Table 3).

Generally, as stocking density increased, lamb production metrics, including number of days to reach target LW, percent of lambs sold at target LW and carcass weight decreased, but total lamb production/ha increased (Table 3). Ewe condition at joining and reproduction in their second lambing cycle also declined as stocking density increased (Table 4). This was because the on-farm feedbase was unable to sufficiently support livestock demand at higher stocking densities, and the maintenance level supplementation provided was inadequate to support high growth rates. Spring-lambing systems were especially sensitive to increased stocking density due to the much higher feed deficits (thus supplement demand) over summer/early autumn compared with the autumn-lambing systems. Although total lamb production increased because of the higher lamb numbers per hectare, the relative increase in total lamb production was marginally lower for spring-lambing systems compared with autumn-lambing systems (17% vs 21%; Table 3).

At all locations and stocking densities, systems that integrated dual-purpose crops generally had higher gross margins than the pasture-only systems due to economic grain yield, and were advantageous across a broad range of seasonal conditions. However, the degree of these differences varied between locations (Fig. 4), with economic grain yield higher on average at the higher-rainfall, cooler environments of Armidale (AUD$293/ha) and Goulburn (AUD$225/ha) compared with the lower-rainfall, warmer environments of Gulargambone (AUD$104/ha) and Temora (AUD$153/ha), irrespective of stocking density. When compared with the standard pasture-only spring-lambing system, the average median gross margins of autumn- and spring-lambing systems with dual-purpose crops were AUD$284/ha and AUD$260/ha higher, respectively (Table 5). As lamb production was generally similar between the two feedbase systems for the same lambing time, these differences were mostly attributable to the additional income from grain harvested from the dual-purpose crops, along with the lower supplement demand of these systems, especially when autumn-lambing (Table 2; Fig. 3). Overall, dual-purpose crops allowed for an increase in stocking density and increased average gross margin without increasing risk in poorer seasons, especially in autumn-lambing systems in higher-rainfall, cooler environments (Table 5).

Fig. 4.  Cumulative distribution of system gross margins for spring- (dotted) and autumn- (solid) lambing systems with pasture-only (green) or with the integration of dual-purpose crops (DPC; gold) at ‘low’, ‘medium’ or ‘high’ stocking densities (30%, 40% and 50% feedbase utilisation rates, respectively) at Armidale (a), Goulburn (b), Gulargambone (c) and Temora (d), NSW. Results represent the annualised gross margin accumulated over each 2-year simulation cycle.

Table 5.  Simulated annual gross margins for spring- and autumn-lambing systems with pasture-only or with the integration of dual-purpose crops (DPC) stocked at ‘low’, ‘medium’ and ‘high’ stocking densities (30%, 40% and 50% feedbase utilisation rates, respectively) at four locations in NSW.

The lambing time and feedbase system that returned the greatest system gross margins differed between locations and stocking densities. In higher-rainfall, cooler environments of Armidale and Goulburn, there was significant benefit to integrating dual-purpose crops compared with pasture-only systems. This increase in gross margins was greatest for autumn-lambing systems, especially at high stocking density where dual-purpose crops reduced supplement demand the most (Table 5; Fig. 4a and b). At these two locations, the autumn-lambing system with dual-purpose crops also carried significantly less risk, returning AUD$386/ha in the lowest 20% of years (i.e. poorest seasons) compared with the spring-lambing system with dual-purpose crops that returned AUD$315/ha, and both pasture-only systems that returned on average ~AUD$183/ha and were often negative at Goulburn (Table 5).

At lower-rainfall, warmer environments of Gulargambone and Temora, systems with dual-purpose crops returned the highest gross margins, but this was for spring-lambing systems rather than autumn-lambing systems (as above; Table 5; Fig. 4c and d). At Temora, spring-lambing systems with dual-purpose crops always returned the highest gross margins, regardless of stocking density. At Gulargambone, the gross margin advantage of dual-purpose crops occurred only at low and medium stocking densities, but was reduced at the highest stocking density (Table 5; Fig. 4c). At Temora, lambing systems with dual-purpose crops always carried significantly less risk compared with the pasture-only systems, especially at low and medium stocking density, returning on average ~AUD$77/ha more in the lowest 20% of years (Fig. 4d). However, at Gulargambone, systems with dual-purpose crops did not provide a significant risk benefit compared with the pasture-only spring-lambing system, returning on average ~AUD$338/ha and AUD$372/ha, respectively, in the lowest 20% of years. This difference was consistent across the different stocking densities (Table 5; Fig. 4c). This was due to failed crops at this location in the driest seasons, which limited winter feed and grain yield, reducing the profit advantages of dual-purpose crops under poor seasonal conditions.

The most important finding of this study is that integrating dual-purpose crops into the feedbase can facilitate earlier lambing in autumn across different environments, because they help to fill the winter feed gap and reduce feed deficit risk (both frequency and intensity), which currently constrain pasture-only autumn-lambing systems. Although dual-purpose crops had little or no impact on lamb production outcomes in both spring- and autumn-lambing systems, additional income from grain yield and lower supplementation of livestock resulted in higher gross margins compared with a pasture-only feedbase. Furthermore, integrating dual-purpose crops into the whole-farm system reduced economic risk, particularly in more marginal years. A shift from spring- to autumn-lambing systems also had positive implications to ewe reproductive performance in the subsequent year, as facilitated by dual-purpose crop integration. We showed that combining dual-purpose crops with autumn-lambing systems can achieve significant advantages in higher-rainfall cooler environments within both summer- and uniform-rainfall zones. We also showed that dual-purpose crops can be beneficial in warmer and drier environments where spring-lambing systems are still more advantageous.

This paper expands on previous work to show that the value of dual-purpose crops differs between environments due to a difference in the combination of livestock demand and local forage supply. Previous livestock modelling for a single location (Wagga Wagga, NSW, Australia) showed that allowing livestock to access dual-purpose wheat as additional grazing land over winter helped to mitigate the winter feed gap and enabled for the safe increase in stocking rates over winter without increasing supplementary feed. In that same modelling study, June lambing systems at eight breeding ewes/ha that integrated dual-purpose wheat had the highest gross margins; however, gross margins were only marginally higher than lambing in May at eight and 10 breeding ewes/ha with the same feedbase (McGrath et al. 2014). Unlike the McGrath et al. (2014) study, we showed that in systems where dual-purpose crops replace 25% of the farms pasture area, the benefits are highly variable depending on the climatic conditions. The main drivers of this variability are largely due to differences in the underlying on-farm feedbase, seasonal feed supply and stocking density, which in turn influence optimal lambing time. For example, the higher-rainfall, cooler locations of Armidale and Goulburn have significant feed deficits in autumn and winter, but abundant feed in spring/early summer. Hence, dual-purpose crop integration is highly beneficial in these locations, because they provide improved feed supply over winter. We showed in our study that providing this additional winter feed can support earlier lambing in autumn and, thus, sale of livestock in late spring/early summer and/or increases in stocking density of the enterprise without dramatically increasing the risk of feed deficits. Similar benefits of dual-purpose crops have been demonstrated in similar environments in both experimental and modelling studies. In southern Tablelands NSW (similar climate to Goulburn), integration of 1:1 dual-purpose wheat and canola in a spring-lambing system reduced supplementation requirements of ewes over winter compared with a pasture-only system (McGrath et al. 2021a). However, in all experimental years, crops were available for grazing in May that might have supported an earlier lambing system in autumn (McGrath et al. 2021a). Modelling studies have also demonstrated the potential for dual-purpose crops to supply ample winter feed for livestock in the northern Tablelands (Armidale, NSW, Australia); although these were limited to biomass-only estimates over winter (Bell et al. 2012).

In contrast to these sites, at the lower-rainfall, warmer environments, the benefits of the winter feed from dual-purpose crops were less clear. At Gulargambone with higher summer rainfall, late summer/early autumn feed is more limiting, but summer active pastures (e.g. lucerne) provide more substantial feed for livestock over late spring and summer. The displacement of higher production pastures (25% of farm area) with dual-purpose crops shifts feed deficits to summer and early autumn when cropping land is unavailable for grazing and there is insufficient pasture to meet livestock demand. At this location, spring-lambing systems with pasture-only are more favourable, although differences become less clear at higher stocking density. At Temora in the uniform rainfall zone, feed deficits are less severe over summer, and dual-purpose crops help to mitigate feed deficits in winter and early spring. Summer active pastures still provide adequate feed for livestock over spring and summer to support spring-lambing systems, whereas dual-purpose crops help with deferring pastures, so more feed is available in spring. As a result, spring-lambing systems with dual-purpose crops are more often favoured at this location, with no negative impact at higher stocking density. These findings together show that across a broad range of environments, the value of dual-purpose crop integration in livestock systems is not simple, and interacts strongly with the local feed supply and the demand of the livestock enterprise.

Because the local context influences the value provided by dual-purpose crops, there are still other aspects that need to be further explored to better tailor their use across environments and production systems. First, we assumed in our study that 25% of the total farm area would be removed from highly productive pasture (LMU1) and replaced with dual-purpose crops. However, the findings from our work suggest that this assumption may not be widely applicable at all locations. For example, most farms at locations such as Gulargambone and Temora within the mixed farming zone include separate land area for grain crops and pastures. Thus, integration of dual-purpose crops at these locations is more likely to be via replacement of grain-only crops rather than via the displacement of pasture, thus increasing the grazeable area for livestock over the autumn/winter period, a similar system to that modelled by McGrath et al. (2014). Thus, although the productivity and profitability of the system per grazed hectare will be very similar, the ‘crop penalty’ identified at these sites over the spring, summer and early autumn in our study are unlikely to be as acute in practice, because crops are unlikely to displace higher productive pastures. Conversely, in livestock-dominated systems, such as Armidale and Goulburn that reside within the Tablelands region, suitable cropping area is limited (Virgona et al. 2006) and dual-purpose crops would inevitably replace pastures grown on more productive land. For these locations, the 25% proportion of cropped land used in our study, may be realistic estimates to optimise the feedbase and total farm profits. Results from experimental studies at a single location in the southern Tablelands, NSW (Hall, ACT, Australia), have shown that allocating 33% of land to 1:1 dual-purpose wheat and canola crops placed greater pressure on permanent pastures, especially in late summer and autumn, resulting in higher supplementation of livestock during these months (McGrath et al. 2021a, 2021b). Modelling studies for the same location in the southern Tablelands, NSW, suggest allocating 10–20% of land to 1:1 of dual-purpose wheat and canola is optimal for improved farm profit margins (Bell et al. 2015a). Others have suggested a maximum 20% allocation of land to dual-purpose crops for the southern Tableland sheep systems in order to maximise profits with lower financial risk (McGrath et al. 2018; 2021c). It is clear from the results of our study and results from others that the optimal area to allocate to dual-purpose crops is location specific, and likely to differ across a broad range of environments and production systems.

Second, early and timely sowing of dual-purpose crops is critical to maximise grazing opportunities. Generally, dual-purpose crops are sown before the end of April, and grazing typically commences in June. In the southern Tablelands NSW (e.g. Goulburn), sowing by 7 March can provide an additional 3–5 weeks of grazing in late autumn, a period where pasture production is generally low, and a ‘crop penalty’ arises (McGrath et al. 2021c). In our study, only dual-purpose crops grown at Goulburn, of similar climatic conditions to the location in the McGrath et al. (2021a, 2021b) study, were sown on average by early March. Sowing dates at all other sites were ~3–4 weeks later, and the sowing of wheat at Temora was nearly 8 weeks later. The sowing of the dual-purpose crops in our modelling study were dictated by a set of sowing rules that only permitted sowing when 25 mm of rain fell over a minimum of 3 days. However, in practice, producers are more likely to sow following the first opening of rains, and in drier environments such as Gulargambone and Temora, that is most likely to occur after a smaller scale rain event or when rainfall exceeds pan evaporation over a 7-day period (Unkovich 2010). Furthermore, the crop grazing rules in our simulations that dictated when the crop could be grazed based on a biomass threshold may have limited grazing opportunities, especially in the lower-rainfall environments. In practice, producers may commence grazing sooner than our biomass thresholds in our simulations permitted, especially when late autumn feed is limited. Theoretically, grazing is safe to commence once the crops have grown secondary roots and are well anchored, which is usually the three leaf stage for cereals and six to eight leaf stage for canola (Nicholson et al. 2016). Dual-purpose wheat grazing studies in Wagga Wagga, NSW, commenced grazing in mid to late June, with levels as low as 0.6–1.1 t DM/ha (McGrath et al. 2015). Designing a flexible grazing rule that can initiate earlier grazing when appropriate might enable a better representation of the potential grazing value of the dual-purpose crops across environments; however, there are significant hurdles in adding this level of complexity into a whole-farm systems model.

Finally, the mix of forages within the farm feedbase is likely to vary considerably between farms at any given location, which will influence the value of dual-purpose crops to fill feed gaps at critical times of the year. A changing climate is also likely to impact the abundance or frequency of early biomass from dual-purpose crops for grazing. Although many livestock systems rely predominately on perennial grass-based pastures (as assumed in our current analysis), mixed farming systems have the opportunity to diversify their forage sources to strategically design feed systems that mitigate climatic risks. This could involve the use of alternative forage legumes, such as deep-rooted annual regenerative pasture legumes (e.g. French serradella (Ornithopus sativus Savi), biserrula (Biserrula pelecinus L.) or bladder clover (Trifolium spumosum L.)), to fill late autumn and winter feed gaps (Hackney et al. 2012), autumn-sown forage brassicas to fill winter and summer feed gaps (Bell et al. 2020; Watt et al. 2021), and improved multispecies pastures to enhance feed supply at different times of the year (Daly et al. 1996; Cranston et al. 2015). Better understanding of the whole-farm feedbase and patterns of seasonal supply is critical for the successful integration of dual-purpose crops to maximise their utilisation over the late autumn and winter, but mitigate their negative implications in late summer and early autumn, when the crop is not yet established or available for grazing.

Although dual-purpose crops may allow for autumn-lambing systems via the provision of autumn/winter feed for late pregnant/lactating ewes, our modelling study showed that they made little or no improvement in lamb production metrics in both spring- and autumn-lambing systems. This is because lambs have very little or no grazing of the dual-purpose crops, because autumn lambs are very young when ewes are grazing the crop, and crop grazing has ceased by the time spring lambs are born. Furthermore, both sets of offspring are sold before the next crop is available for grazing the following season. Because lamb production was largely dependent on the quantity and nutritive value of the pasture available to them after weaning, any differences in lamb production that did arise generally favoured the pasture-only system, because the dual-purpose crops displaced the most productive pastures. Thus, differences in gross margins between the pasture-only and dual-purpose crop systems can only be attributed to differences in supplementary feeding expenses and economic grain yield, which generally favoured the dual-purpose crop systems.

Although there were no substantial differences in lamb production between the two feedbase systems, shifting lambing time from spring to autumn had a positive impact on individual lamb production metrics. We found autumn-born lambs weaned onto early spring pastures grew faster (and were sold sooner) compared with spring-born lambs weaned onto late spring/early summer pastures. Furthermore, grazing of highly productive spring pastures from weaning to sale allowed autumn-born lambs to maintain a constant plane of nutrition that reduced supplementary feeding compared with the spring-born lambs, whose plane of nutrition fluctuated with seasonal deficits (over summer and early autumn). As a result, a much higher proportion of autumn-born lambs were able to meet target LW, leading to much higher average carcass weights compared with spring-born lambs. Although the breakdown of gross margins is not specifically reported in the results of this study, lamb sale estimates indicate that autumn-born lambs with heavier carcass weights achieved on average ~$13/lamb more than spring-born lambs, with differences most apparent in higher-rainfall, cooler environments. However, sheep are seasonal breeders, and in temperate environments, ewe ovulation and oestrus peak in autumn, when spring-lambing ewes are joined (Rosa and Bryant 2003); thus, out-of-season joining can reduce lambing rates. In our modelling study, spring-lambing systems had ~13% higher weaning rates than autumn-lambing systems, resulting in marginally higher total lamb production and ~3% higher total lamb income compared with autumn-lambing systems. To maximise the benefits of autumn-lambing systems, but mitigate lower weaning rates from out-of-season breeding, artificial methods to improve reproductive rates can be administered, including melatonin (e.g. Regulin®) or a progesterone implant (e.g. CIDR®; Fisher 2004). Interventions with these technologies may further improve the benefits of an autumn-lambing system and total lamb production, and overall farm profits. However, a system with more lambs may also present its own set of challenges, and an increase in weaning rates in autumn-lambing systems should only be considered where lamb production benefits outweigh the cost of supplement required to meet target LW late in the season, when feed quantity and nutritive value is in decline.

A particularly novel component of this modelling study was the beneficial legacy effects of an autumn-lambing system on ewe body condition at joining and subsequent reproductive performance (i.e. based on lamb weaning rates) in the following year, which was facilitated by the integration of dual-purpose crops. We found there was little difference in ewe body condition or reproductive performance between the different feedbase systems, which concurs with experimental findings (McGrath et al. 2021a). In our modelling study, this was because ewe body condition was maintained via supplementary feeding, with the frequency of feeding dictated by a minimum ewe BCS of 2.5 (based on the average of the ewe flock), which came at a high cost for autumn-lambing systems without dual-purpose crops. However, we did find that autumn-lambing ewes had greater capacity to improve body condition after weaning compared with spring-lambing ewes, which led to increased reproductive performance (i.e. lamb weaning rates) in the second lambing cycle. Autumn-lambing ewes were better able to maintain a high plane of nutrition from weaning to joining when spring pastures were at their peak. In contrast, spring-lambing ewes were lactating during the peak of spring pastures, which was followed by feed deficits and lower quality forage over the summer months before being joined again in early autumn. As a result, ewes from spring-lambing systems often had a BCS < 3.0, which is below the recommended score for joining and pre-lambing ewes (Australian Wool Innovation and Meat & Livestock Australia 2008).

Mixed farming systems enable livestock and cropping enterprises to co-exist, while spreading economic risk between enterprises. Our modelling study showed that integrating dual-purpose crops can significantly improve farm gross margins in either spring- or autumn-lambing systems, with benefits arising from fewer supplementary feed expenses over winter, especially in autumn-lambing systems. Dual-purpose crops also contributed significant economic grain yield, while livestock production was relatively consistent. The gross margins estimated in our study are more robust compared with other more simple gross margin analyses (McGrath et al. 2014), having considered key input costs of the whole-farm system, including livestock input costs (e.g. shearing, crutching, replacement ewes purchases, labour, livestock infrastructure and pasture management) and crop input costs (e.g. seed, fertiliser, pesticides and contractors), and the major income streams of both enterprises (e.g. lamb and sheep meat, wool, and grain production).

In higher-rainfall, cooler locations (i.e. Armidale and Goulburn), gross margins were highest for systems that integrated dual-purpose crops, and when combined with autumn-lambing, carried significantly less risk in poor seasonal conditions than spring-lambing systems, especially at higher stocking density. This was largely attributable to the much higher supplementary feed requirements of spring-lambing systems in late summer and autumn in poorer years, which was exacerbated by the displacement of productive pastures for crops. In lower-rainfall, warmer locations (Gulargambone and Temora), gross margins were highest for systems that integrated dual-purpose crops, but differences between lambing systems were marginal. Furthermore, the benefit of dual-purpose crops in mitigating economic risk in poor seasonal conditions was less clear, especially at Gulargambone. At Gulargambone, lambing systems that integrated dual-purpose crops and the spring-lambing system with pasture-only carried similar economic risk in poor seasonal conditions. This was because the displacement of pasture for crops in late summer and early autumn led to higher supplementary feed expenses; however, this was often offset by economic grain yield. At Temora, integrating dual purpose crops in either autumn- or spring-lambing systems reduced economic risk in poor seasonal conditions, but spring-lambing systems with dual-purpose crops were marginally more favourable. At higher stocking density in poor seasonal conditions, the difference between all the scenarios tested was marginal, as systems with dual-purpose crops had higher supplementary feed expenses than at lower stocking densities, combined with the generally lower grain yield in those poorer years. At both Temora and Gulargambone, the greatest benefit of dual-purpose crops relative to the pasture-only systems was under better seasonal conditions, even under higher stocking densities, as supplementary feed requirements were less, and grain yield was higher.

Although our gross margin estimates considered multiple input costs and income streams of the system, there were some considerations that influenced our analysis. First, there are likely to be regional differences in costs associated with crop protection and fertiliser inputs, and temporal variations in livestock prices that may alter the gross margins obtained; they were maintained consistent across sites here to enable direct comparisons. Second, in areas where cropping is not a standard practice, such as Armidale and Goulburn, contractors may be more difficult to acquire, and producers may overcapitalise on cropping equipment. Furthermore, some livestock producers are skilled in the management of livestock, but lack knowledge in cropping systems, which presents an additional challenge when implementing these system changes (Francis 2022). Better understanding of all the factors that influence the viability and economic performance of dual-purpose crops is fundamental to successful application on-farm, and further research is warranted to better understand these potential constraints in new areas.

Our simulation modelling demonstrates that across different environments, integrating dual-purpose crops helps to fill the winter feed gap and facilitates autumn-lambing systems. Compared with the standard pasture-only spring lambing system, integrating dual-purpose crops both in spring- and autumn-lambing systems improved total farm gross margins due to economic grain yield and reduced supplementary feed. Furthermore, integration of dual-purpose crops supported higher stocking densities and lowered economic risk, whereas a shift from spring- to autumn-lambing facilitated by dual-purpose crops also led to better reproductive performance of ewes in the subsequent year. In higher-rainfall, cooler environments, autumn-born lambs were sold in early summer, which limited the impact of the summer forage deficit, and could have additional benefits for animal health and marketing, which warrants further exploration. Although there are limitations to modelling studies, we showed that whole-farm systems modelling can provide considerable insight into the complexity of dual-purpose crop integration across environments using multi-year simulations, which is difficult to achieve with experimental studies.

Supplementary material is available online.

The data that support this study will be shared upon reasonable request to the corresponding author.

The authors declare no conflicts of interest.

This paper was prepared during a Meat and Livestock Australia Donor Co. project (P.PSH.1045: Dual-purpose crops for lamb production in southern Qld and northern NSW).