Cereals after pasture legumes have higher grain protein levels

Key messages

  • The Dryland Pasture Legume Systems Project has investigated the impact of new pasture legumes on subsequent cereal crops.
  • Cereals following pasture legumes in a rotation have higher grain protein levels.
  • Cereal yield was the most responsive in the year after fallow


Determine if biologically fixed nitrogen from elite pasture legumes can increase grain protein levels in dryland farming systems across multiple soil types in Western Australia.


Large quantities of Western Australia’s cereal crops are not achieving grain protein levels preferred by markets. The main reasons include; forage legume density in pasture paddocks have declined, the soils are generally lower in organic matter and nutrients, and canola has become a profitable break crop (McKenzie 2018). Further, there is little incentive for farmers to make the wheat grade of APW1 (10.5% protein) at the receival point, yet there are larger discounts for low protein levels than premiums above the critical value. Continuous cropping systems in WA (without legumes) are heavily reliant on applications of fertiliser nitrogen (N) to optimise yield and protein. The N availability in WA soils is often low and difficult to predict as stable organic N pools are highly transitory (Moore 2001). Optimising grain yield and protein requires the farmer to predict the seasonal growing conditions and N mineralisation rates so that N supply can be matched to crop demand. As climate systems become increasingly variable this prediction becomes less reliable.

Interpreting N transformations can be as simple as agronomically diagnosing the season’s results after harvest (Bestow 1992). The simple scenarios are:

  • High protein and yield is associated with sufficient N and rainfall;
  • Low yield and high protein is associated with sufficient N and insufficient rainfall;
  • High yield and low protein is associated with insufficient N and sufficient rainfall, diluting the N for protein;
  • Low yield and protein is associated with insufficient N and rainfall.

When synthetic N has been historically cheap the logical management strategy has been to hedge on the oversupply of N. Therefore, in years of good rainfall increased yield potential can be captured while in low rainfall years farmers may be able to capture higher grain protein opportunities. However, as the cost of N increases (in both economic and environmental sense), the logic may shift to an undersupply which will reduce the potential to optimise production.

Biological N2 fixation from legumes in a rotation with cereals can significantly increase the sustainability of a system (Crews and Peoples 2005, Kirkegaard et al. 2008). Although pulse legumes provide a profitable break, the majority of the above ground N that is fixed is subsequently removed in the grain harvest (Peoples et al. 2001). While plant biomass (and therefore N) is removed via livestock production, pasture legume systems demonstrate a greater ability to fix and retain N in the system (Fillery 2001, Peoples and Baldock 2001). Recent development of new cultivars for dryland systems (UMU1805-001RMX) and new establishment techniques such as summer sowing (Nutt et al 2021) have indicated greater amounts of nitrogen fixation. However, the rate of N fixation, and the impact on cereal quality for dryland agricultural systems was yet to be evaluated. This study examined the impact of pasture legumes on the yield and protein content of subsequent cereal crops grown in dryland cropping systems of WA.


Three randomised and replicated (n=4) crop rotation experiments were undertaken in Ardath, Narembeen and Canna. The soil type at Ardath was coarse textured acidic sand while Narembeen and Canna were fine textured loams. The legumes that were sown differed between sites to utilise the best cultivar for the soil type and rainfall (Table 1).


Varieties sown

The cereal varieties (barley and/or wheat) sown each year (in order of year) were as follows: Ardath= cvs. Spartacus CL, Chief CL, Maximus CL and Sceptre; Narembeen= Spartacus CL, Sceptre, Sceptre; Canna= Maximus CL, Sceptre.

Soil testing

Soil was sampled at the start of every growing season using a stainless-steel corer, with samples in every plot taken in increments of 10cm. Soil analysis was undertaken by CSBP (Bibra Lake), and there was no significant difference in the soil total nitrogen between plots within sites except for the fallow plots being significantly higher.

Rotations and biomass

Rotations are described in Tables 2 to 4. In the seasons where every rotation was in the same cereal crop (wheat or barley), the plots were sown by the host farmer with broadacre equipment. During the season when the legumes were sown or regenerated, the legumes were subsampled within the plot using a quadrat to estimate peak biomass (t DM/ha). Legume biomass was measured in spring by cutting four random quadrats (0.1m2) per plot which were then dried and weighed to estimate t DM/ha of above ground biomass. The serradella, trigonella and medic plots were ungrazed throughout the experiment, however the vetch was harvested at the end of the season.

Each legume or barley plot sown in the first year was sown using an Aitchison mini seeder with five passes per plot in the same direction (to give 5m width). Plots in Ardath were 20m long while plots in Canna and Narembeen were 10m long. In the cereal cropping phase, the 10m long plots were split into three sub plots (3.3m x 5m) or the 20m plots split into four sub plots (5m x 5m). Each plot that received N was in the form of urea (46% N) at the early tillering stage (Zadock 20-25), 24h before a significant rainfall event (>5mm). At the end of the season, each subplot was harvested using a small plot harvester. The chaff was spread evenly across the subplot to eliminate the chance of uneven nutrient cycling or moisture conservation. Grain yield (t/ha), protein levels (%) and screenings (not presented) data were collected and analysed using ANOVA (strip plot design with complete randomised blocks of the main rotation treatment). The least significant difference (lsd) values were used to compare different rotations within the same N application strategy within each site and year.



Average serradella peak biomass in 2018 was 2.5 (DM t/ha) and the 2020 regenerating serradella plots averaged 2.3 (DM t/ha). Soil N in the legume plots did not exceed 22.3mg/kg. For most treatments across all rotations, the grain protein levels were significantly higher in the legume plots than the continuous cereal control for three years (Table 2). Grain protein levels were also significantly higher after the fallow across all N rates except the 40N/ha rate in 2020. However, in 2021 (the wettest year studied) the ex-serradella plots were the only plots that had significantly higher wheat grain protein levels, but only reached the grade of APW1 at the highest N rate. Wheat yield after regenerating serradella was significantly higher in 2021 in the plots that received the two lowest N rates (7 and 30 N/ha, respectively).



Vetch sown in 2019 produced significantly more biomass (2.5DM t/ha) than the medic (1.3DM t/ha). The high biomass production from the vetch was also associated with high soil nitrate in 2020 (27.3mg/kg) and 2021 (15.3mg/kg). In 2020 (a dry year), the wheat in the ex-fallow and ex-vetch plots significantly outyielded the continuous cropping treatment (Table 3). Furthermore, the 2020 grain protein levels were significantly higher in the ex-vetch plots at all N rates. The 2021 harvest resulted in a significantly higher protein level of 9% at 42N/ha in the ex-vetch plot, however, no grain samples reached the protein requirement for APW1. The break crop effect was more significant in 2020 with significantly higher yields than in 2021 due to the low legume biomass produced in 2019.



In 2020, medic plots (6.3DM t/ha) significantly outyielded trigonella plots (4.4DM t/ha) in peak biomass by 1.9DM t/ha. In the following season, the highest wheat yield recorded was 4.2t/ha from both the ex-medic plots (N rate= 92N/ha) and the ex-trigonella plots (zero N) (Table 4). However, ex-trigonella plots were significantly higher in wheat yield than the continuous cereal control after receiving 0 and 46kg N/ha. All 2020 treatments had significantly higher grain protein than the ex-barley treatment with the highest percentage of 11% attributed to the ex-trigonella plots. When no synthetic fertiliser was applied, the fallow plot had significantly higher protein than either legume treatment.



Global markets are indicating that in the future, low protein grain will be increasingly difficult to sell, which will ultimately lead to disparity between premium and feed grain prices (Lemon 2007). This study has demonstrated that pasture legumes grown across three markedly different climates (across dryland WA) can increase grain protein levels. Also, efficient-fixing pasture legumes (i.e., serradella, vetch and trigonella) used in rotation with cereals can allow farmers to hedge on undersupplying N without significantly compromising yield and protein in high demand years while reducing the chance of oversupply in years of low demand. Compared to other pasture legumes, medics are notoriously poor nodulators when there is a soil type with a large amount of mineralised N (Howieson and Ballard 2004), hence the subsequent wheat crops studied had lower yield and protein levels at Narembeen and Canna. Fallow and deep-rooted legumes have impacts on soil moisture availability in the subsequent season, so there may be trade-offs in crop and pasture sequences to consider.

An important note is that the pasture legumes in this study were not grazed by sheep, it is likely that heavy grazing would redistribute N from the systems studied but would improve the economic returns and potentially offset risk within the system, as seen in related economic modelling of pasture legume effects in rotations (Thomas et al. 2022, GRDC crop updates). Prices of N have increased 175% over the 5-year average (Ostendorf 2021) and it is estimated that synthetic N application can contribute as much as 50% of the CO2 emissions from a farm system (Camargo et al. 2013). If farmers got paid a higher premium for high grain protein content, this could encourage greater use of well-managed pastures, improved soil health, and higher sustainability of the farming system.


The authors acknowledge The Dryland Legumes Pasture Systems project, which is funded by the Australian Government Department of Agriculture Water and the Environment as part of its Rural R&D for Profit program, the Grains Research and Development Corporation, Meat and Livestock Australia and Australian Wool Innovation (Project No. RnD4Profit-16-03-010). We would like to thank the various technicians, grower groups and support staff that have contributed to this project. Finally, we appreciate the input from host farmers Clint Butler (Narembeen), Jeremy Wasley (Canna) and Philip Foss (Ardath). The research undertaken as part of this project is made possible by the significant contributions of growers through both trial cooperation and the support of the GRDC, the authors would like to thank them for their continued support.


Bestow, S. (1992). Quest for quality: operation quality wheat. Perth, Western Australian Department of Agriculture. 6.

Camargo, G. G. T., M. R. Ryan and T. L. Richard (2013). "Energy Use and Greenhouse Gas Emissions from Crop Production Using the Farm Energy Analysis Tool." BioScience63(4): 263-273.

Crews, T. E. and M. B. Peoples (2005). "Can the synchrony of nitrogen supply and crop demand be improved in legume and fertilizer-based agroecosystems? A review." Nutrient cycling in Agroecosystems 72(2): 101-120.

Fillery, I. R. P. (2001). "The fate of biologically fixed nitrogen in legume-based dryland farming systems: a review." Australian Journal of Experimental Agriculture 41(3): 361-381.

Howieson, J. G. and R. Ballard (2004). "Optimising the legume symbiosis in stressful and competitive environments within southern Australia—some contemporary thoughts." Soil Biology and Biochemistry 36(8): 1261-1273.

Kirkegaard, J., O. Christen, J. Krupinsky and D. Layzell (2008). "Break crop benefits in temperate wheat production." Field Crops Research 107(3): 185-195.

Lemon, J. (2007). Nitrogen management for wheat protein and yield in the Esperance port zone. Perth, Western Australia, Department of Primary Industries and Regional Development.

McKenzie, P. (2018). "APW1 as the Benchmark wheat grade in WA." Agricultural Science 29(2/1): 76-81.

Moore, G. A. (2001). Soil Guide: A handbook for understanding and managing agricultural soils. Perth, WA, Western Australian Department of Agriculture.

Nutt, B. J., A. Loi, B. Hackney, R. J. Yates, M. D’Antuono, R. J. Harrison and J. G. Howieson (2021). "“Summer sowing”: A successful innovation to increase the adoption of key species of annual forage legumes for agriculture in Mediterranean and temperate environments." Grass and Forage Science76(1): 93-104.

Ostendorf, M. (2021). What is going on with fertiliser prices? Successful farming. https://www.agriculture.com/news/crops/skyrocketing-fertilizer-market-has-farmers-analysts-and-companies-weighing-in.

Peoples, M. and J. Baldock (2001). "Nitrogen dynamics of pastures: nitrogen fixation inputs, the impact of legumes on soil nitrogen fertility, and the contributions of fixed nitrogen to Australian farming systems." Australian Journal of Experimental Agriculture 41(3): 327-346.

Peoples, M. B., A. M. Bowman, R. R. Gault, D. F. Herridge, M. H. McCallum, K. M. McCormick, R. M. Norton, I. J. Rochester, G. J. Scammell and G. D. Schwenke (2001). "Factors regulating the contributions of fixed nitrogen by pasture and crop legumes to different farming systems of eastern Australia." Plant and Soil 228(1): 29-41.

Contact details

Robert Harrison
CSIRO Agriculture and Food
147 Underwood Ave, Floreat, WA, 6014
Ph: +61 8 9333 6032  
Email: robert.harrison@csiro.au

GRDC Project Code: UMU1805-001RMX,