Impact of crop rotations on profit, nitrogen and ryegrass seed bank in crop sequences in southern NSW

Take home messages

• Profitability of TT-canola > high input wheat > wheat sequence
  • Very profitable if no grass weeds and nitrogen (N) supplied.
  • Not sustainable in the long run.
• Profitability of pulse grain > high input wheat > wheat
  • Close behind on $’s even with low pulse prices.
  • More available soil N for wheat.
• Profitability of RR-canola > high input wheat > wheat
  • Most profitable response to a ryegrass weed problem, but;
  • One year break crop control of ryegrass was found to be insufficient because of high residual ryegrass seed bank.
• Profitability of pulse grain (spray topped) > RR-canola > wheat, or RR-canola > cereal hay > wheat
  • Very profitable.
  • Can provide two years of complete seed set control of ryegrass (but necessary to spray paraquat in pulse or glyphosate before cutting hay crop and paraquat after to achieve complete grass control in years 1 and 2).
• Sequences with brown manures and long fallow are not as profitable, but have other benefits not included in gross margins.
  
  • Can provide two years of complete seed set control of ryegrass (when used as double break with RR-canola), plus;>
  • provide additional N for subsequent crops although strategic N management will be required if following cropping years are dry to avoid haying-off;
  • Improves timeliness of operations
  • Reduces risk and working capital requirements.

Background

There is a wide-spread perception among farmers and their advisers that broadleaf break crops such as canola and legumes are higher risk and not as profitable as cereals. The aim of GRDC project CSP000146 was to challenge this notion, and to examine the potential beneficial impacts of break crops on the longer-term financial performance of following wheat crops. This project recognises that profit and risk are key drivers of farmer decision making and endevours to investigate the overall profitability of the whole cropping sequence. Much of the project’s experimental and communications program is based on the assumption that in the absence of high grain prices for canola or pulses, growers are most likely to want to sow broadleaf break crops to address specific agronomic problems when growing cereals. Consultation with grower group and agribusiness collaborators in project CSP000146 has identified difficulties in managing grass weeds as now one of the main constraints to wheat production across south-east Australia. This paper presents key findings generated from CSIRO experimentation undertaken in association with FarmLink in southern NSW that address three key questions:

1. Can a canola or legume break crop be as profitable as a cereal?
2. Do cropping sequences that include break crops improve the profitability of subsequent cereal crops in cropping systems?
3. Can herbicide resistant ryegrass be managed more cost-effectively under break crops than cereals?

Methodology

Three experiments were established between 2011 and 2013 in paddocks on three different farms in south-eastern NSW. One paddock at Junee Reefs had no herbicide grass weed problem, but the field sites selected on two farms at Eurongilly (Exp 1 and Exp 2) were known to contain herbicide-resistant ryegrass. Various crop sequences were established in each experiment over a three year period with the Junee Reefs experiment concluding in December 2013 and Eurongilly (Exp 1 and Exp 2) concluding in December 2014 and 2015, respectively. The crops / treatments established in the experiments in each of the three years were:

• Year 1: Canola (high or low input), legumes (pulse grain or brown manure), wheat (high or low input), or fallow (Eurongilly exp 1 and exp 2 only);
• Year 2: Wheat (high or low input) and additionally at Eurongilly were canola (Roundup Ready) or wheat (cereal hay);
• Year 3: Wheat – fertilised to achieve a 5t/ha grain yield (if Yield Prophet indicated there was a 70% chance of achieving this when the wheat crop was at growth stage of GS30 of each year).

Part 1: Junee Reefs - Impact of break crops on profit and nitrogen

Aim: To compare the productivity and relative profitability of various low input/low risk cropping options with alternative high input/high risk, but potentially higher return crops.

Methodology – Junee Reefs NSW

The experiment was located in a farmer’s paddock at Junee Reefs located 23km north of Junee. Soil pH (CaCl₂) in the surface 0-10 cm was 5.50. Soil mineral nitrogen (N) prior to the commencment of the experiment in April 2011 (0-150 cm) was 100 kg N/ha.

Junee Reefs (2011) – Initial break crop and cereal treatments

The treatments imposed in 2011 were described in considerable detail in the Update papers by Peoples or Swan in 2012 and 2013, and can be provided upon request from the authors. In summary the range of crops sown, their relative risk category and input costs were as follows:

1. Canola - low: cv Crusher TT open pollinated variety; input costs = $181/ha.
2. Canola - high: cv Hyola® 505 RR Hybrid; input costs = $334/ha.
3. Lentil - medium: cv Flash; input costs = $172/ha.
4. Chickpeas - high: cv Slasher; input costs = $296/ha.
5. Lupins : cv Mandelup for grain; input costs = $164/ha.
6. Lupins BM - low: cv Mandelup for brown manure (BM); input costs = $115/ha.
7. Field peas BM - low: cv Morgan for brown manure (BM); input costs = $104/ha.
8. Wheat - low: cv Lincoln; input costs = $117/ha.
9. Wheat - high: cv Lincoln; input costs = $324/ha.

There were two sowing times: late April (canola and lupins) and mid May (lentil, chickpeas, fieldpeas and wheat). Treatments were replicated four times and were sown in a randomized design in 2.5 x 20m plots.

Junee Reefs (2012) - First wheat after break crops

Each of the replicated plots of all treatments from 2012 were split into 2x10m sub-plots and sown to Spitfire wheat in mid-May 2012 under either a low or high input regime.

Low input wheat: target density of 75 plants/m², seed-dressed and deep-banded starter fertiliser of 25 kg/ha MAP (2.5 kg N/ha), top-dressed with 100 kg/ha urea (46 kg N/ha). Total input cost including initial knock-down, in-crop herbicides and foliar fungicide was $143/ha.

High input wheat: target density of 150 plants/m², seed-dressed and deep-banded starter fertiliser of 75 kg/ha MAP (7.5 kg N/ha), top-dressed with 200 kg/ha urea (92 kg N/ha). Total input cost including initial knock-down, in-crop herbicides and foliar fungicide was $379/ha.

Rationale for inclusion of the choice of various inputs in the high input treatment was:

• Higher seeding rates for greater competition with weeds
• Flutriafol (Impact) to reduce the risk of take-all disease
• Difenoconazole + metalaxyl-M (Dividend) and zinc seed dressing for rhizoctonia
• More starter fertiliser for early vigour for competition with weeds and root diseases
• Use of new pre-emergent grass herbicide chemistry
• Use of new post-emergent broadleaf herbicide chemistry to maintain competition.

Junee Reefs (2013) - Second wheat after break crops

All plots were sown to wheat cv Gauntlet, at a target density of 150 plants/m²; seed dressed with Dividend and Activist Zinc and fertilised with MAP dressed with Impact Endure banded with the seed at 50 kg/ha (5 kg N/ha). All plots had an initial knockdown of Roundup Attack plus Goal, pre-emergent herbicides of Logran 750 WG and Avadex Xtra with in-crop pesticides being Velocity, Prosaro and Pirimor WG insecticide. The difference in total input costs between the 2013 treatments relates to the amount of urea applied with 200 kg/ha (92 kg N/ha) spread onto the wheat and canola 2011 treatments, 100 kg/ha (46 kg N/ha) onto the chickpea and lentil treatment and 50 kg/ha (23 kg N/ha) applied to the fieldpea BM and lupins grain treatments. No urea was applied to the lupins BM 2011 treatment. Total input costs for each group of treatments were:

1. Wheat, canola, barley = $345/ha.
2. Chickpeas and lentils = $280/ha
3. FieldpeasBM and lupins grain = $245/ha
4. LupinBM = $215/ha

Part 2: Eurongilly - Impact of break crops on weeds and profit

There is substantial evidence from surveys of the frequency of herbicide resistance in weeds collected at random from farmers’ paddocks in southern NSW, SA and Victoria that have indicated there is wide-spread resistance or partical resistance to a wide range of herbicide groups (up to 70-80% of samples in some areas; Broster et al 2011; Preston et al 2013). One of the most common reasons farmers give for replacing a cereal with canola or legume is to provide new herbicide options to manage difficult to control weeds. Two studies were initiated at Eurongilly in southern NSW, one in 2012 (Exp 1) and one in 2013 Exp 2) to investigate the efficacy of different break crop strategies and herbicides.

Aim

To compare the productivity and profitability of cropping in the presence of a background herbicide-resistant annual ryegrass population, and to assess the implications of various low or high input grass weed control options applied to wheat and break crops on production costs, ryegrass management and N.

Methodology

The susceptibility / resistance of the annual ryegrass populations at both the Eurongilly trial sites (Exp’s 1 and 2) were tested by Plant Science Consulting SA in March 2012 and 2013, respectively. The result of the analysis indicated that the ryegrass was resistant to Group A herbicides Verdict (Haloxyfop), Select (Clethodim) and Axial (Pinoxaden & Cloquintocet-mexyl) and Group B herbicides, Hussar (Iodosulfuron-methyl-sodium) and Intervix (Imazamox + Imazapyr) to varying degrees (30% to 95%), but suggested that the ryegrass was still susceptible to Group M (glyphosate) and one Group A herbicide, Factor (Butroxydim) (Table 1). This information was used to determine the research herbicide treatments for experimental purposes.

Table 1: Herbicide resistance assessment of annual ryegrass populations in Exp 1 and Exp 2 at Eurongilly, NSW (determined in March 2012 and March 2013).

Resistance- rating: RRR - indicates plants tested have strong resistance; RR- indicates medium- level resistance; R - indicates low- level but detectable resistance; S - indicates no detection of resistance.

^{ɸ a} Hussar label recommends 0.25% BS1000; ^{ɸ b} Intervix label recommends Hasten at 0.5%

In late March 2012 (Exp 1) and 2013 (Exp 2), the initial annual ryegrass seed bank population was determined by removing forty 5cm deep x 6cm diameter surface soil cores from across each experimental site. The soil was put into trays and watered over the next three months and all ryegrass plants that germinated and emerged were counted. The initial ryegrass seed bank in each experiment was:

• Experiment 1 (2012): 1815 plants/m²
• Experiment 2 (2013): 2775 ryegrass plants/m².

The soils at both experimental sites were red chromosols, (Isbell 1996). At the start of the experiments, the soil pH (CaCl₂) of the surface soil in Experiment 1 was 4.5 at 0-10 cm and 10-20 cm prior to applying lime at 2.5 t/ha in April 2012, and pH (CaCl₂) 5.3 and 4.6 (0-10cm and 10-20cm) in experiment 2 in April 2013. Olsen available P (0-10 cm) was 29 mg/kg (Exp 1) and 44 mg/kg (Exp 2), and soil mineral N (0-150 cm) was 87 kg N/ha (Exp 1) in 2012 and 90 kgN/ha (Exp 2) in 2013 just prior to sowing. In March, year 1 (pre-experiment) in experiments 1 and 2, forty soil cores (6cm in diameter x 5cm deep) were randomly removed across the trial area with eight surface cores removed per treatment in April of year 2, 3 and 4 to measure changes in ARG seed bank numbers.

Monthly, annual and growing season rainfall at Eurongilly is outlined in Table 2. Growing season rainfall (GSR, Apr-Oct) in 2012, 2013 and 2014 at Eurongilly were 179 mm, 274 mm and 246 mm respectively, compared to a long term average of 328 mm. Annual rainfall (AR) in the three years was 557 mm, 432 mm and 456 mm respectively, compared to long-term average annual rainfall (AAR) of 556 mm.

Table 2: Monthly, yearly (AR) and growing season rainfall (GSR) in 2012, 2013 and 2014 at Eurongilly, NSW.

Year 1 treatments imposed, input/risk categories and input costs for Exp1 (2012) and Exp 2 (2013) at Eurongilly:

1. Canola (L) – low: cv’s Crusher TT (Exp 1) and Stingray TT^A (Exp 2), target density of 40 plants/m², seed dressed with Jockey Stayer and Gaucho, sown with MAP (25 kg/ha), then topdressed with ammonium sulphate (100 kg/ha) and urea (100 kg/ha Exp 1 and 160 kg/ha Exp 2). Total fertiliser elements as N:P:S (kg/ha) were 69.5, 5.5 and 4.4 (Exp 1) and 97.5, 5.5 and 24.4 (Exp 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent of 480 g/L trifluralin @ 2 L/ha; 900 g/kg atrazine @ 1.1 kg/ha; in-crop herbicide 250 g/kg butroxydim @ 80 g/ha + 900 g/kg atrazine @ 0.9 kg/ha (Exp 1) or clethodim @ 500ml/ha + 900 g/kg atrazine @ 0.9 kg/ha (Exp 2). Total input costs for seed, fertiliser, herbicides and insecticides = $249/ha and $305/ha (Exp’s 1 and 2).
2. Canola (H) – high: cv Hyola® 505 RR Hybrid (Exp’s 1 and 2), target density of 40 plants/m ², seed dressed with Jockey Stayer and Gaucho, sown with MAP (75 kg/ha) and Impact Endure, then topdressed with ammonium sulphate (100 kg/ha) and urea (200 kg/ha Exp 1 and 360 kg/ha Exp 2). Total fertiliser elements as N:P:S (kg/ha) were 120.5, 16.4 and 25.1 (Exp 1) and 195.5, 16.4 and 25.1 (Exp 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent of 480 g/L trifluralin @ 2 L/ha; in-crop herbicide 690 g/kg glyphosate (Round-Up Ready) @ 0.9 kg/ha at 2-3 leaf and 6 leaf stages. Total input costs for seed, fertiliser, herbicides and insecticides = $427/ha and $524/ha (Exp’s 1 and 2).
3. Fallow - low: Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; fallow established in September 2012 with an application of 450 g/L glyphosate @ 2 L/ha + metsulfuron-methyl @ 5 g/ha, then follow-up with 250 g/L paraquat @ 2 L/ha. Total input costs for herbicides = $35/ha and $57/ha (Exp’s 1 and 2).
4. Field peas BM - low: cv’s Morgan (Exp 1) and Percy (Exp 2) for brown manure (BM), target density of 40 plants/m², sown with MAP (25 kg/ha). Total fertiliser elements as N:P:S (kg/ha) were 2.5, 5.5 and 0.4 (Exp’s 1 and 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent 480 g/L trifluralin @ 2 L/ha; 900 g/kg simazine^ɸa @ 1.0 kg/ha; brown manure herbicide 450 g/L glyphosate @ 2 L/ha + 300 g/L clopyralid @ 150 ml/ha + 240 g/L carfentrazone-ethyl^ɸb @ 25 ml/ha; fallow maintenance 450 g/L glyphosate @ 2.5 L/ha. Total input costs for seed, fertiliser and herbicides = $120/ha and $169/ha (Exp’s 1 and 2). (n.b. ^{ɸ a}not registered in field peas; ^ɸbnot registered for this use)
5. Lupin grain - high: cv Mandelup for grain (Exp’s 1 and 2) target density of 40 plants/m², sown with MAP (75 & 25 kg/ha) Exp’s 1 & 2. Total fertiliser elements as N:P:S (kg/ha) were 7.5, 16.4 and 1.1 (Exp 1) and 2.5, 5.5, 0.4 (Exp 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent 480 g/L trifluralin @ 2L/ha; 900 g/kg simazine @ 2.2 kg/ha; in-crop herbicide 250 g/kg butroxydim @ 180 g/ha (Exp 1) or clethodim @ 500ml/ha (Exp 2); 250 g/L paraquat @ 400 ml/ha spray topped mid November (Exp 1 only). Total input costs for seed, fertiliser & herbicides = $168/ha & $161/ha (Exp’s 1 and 2).
6. Wheat (L) - low: cv’s Spitfire (Exp 1) and Gauntlet (Exp 2), target density 75 plants/m², seed dressed with Raxil, sown with MAP (25 kg/ha) then topdressed with urea (100 kg/ha). Total fertiliser elements as N:P:S (kg/ha) were 48.5, 5.5, 0 (Exp’s 1 and 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent 480 g/L trifluralin @ 2 L/ha + diuron 500 g/L @ 1 L/ha; in-crop herbicide 800 g/L prosulfocarb +120 g/L s-metalochlor (Boxer Gold) @ 1.5 L/ha at 2-3 leaf stage^ɸa. Total input costs for seed, fertiliser and in-crop herbicides and foliar fungicide = $169/ha and $164/ha (Exp’s 1 and 2). (n.b. ^{ɸ a}label recommends only pre-emergent use).
7. Wheat (H) - high: cv’s Spitfire, (Exp 1) and Gauntlet (Exp 2), target density 150 plants/m², seed dressed with Dividend, sown with MAP (75 kg/ha) and Impact Endure, then topdressed with urea (200 kg/ha Exp 1 and 360 kg/ha Exp 2). Total fertiliser elements as N:P:S (kg/ha) were 99.5, 16.4, 0 (Exp 1) and 173.1, 16.4, 0 (Exp 2). Herbicides for ryegrass control included initial knock-down with 450 g/L glyphosate @ 1.6 L/ha; pre-emergent 850 g/kg pyroxasulfone (Sakura 850WG) @ 118 g/ha + 500 g/L tri-allate (Avadex Xtra) @ 2 L/ha; in-crop herbicide 800 g/L prosulfocarb +120 g/L s-metalochlor (Boxer Gold^ɸa) @ 2.5 L/ha + 100 g/L pinoxaden 25 g/L cloquintocet-mexyl (Axial) @ 150^ɸb ml/ha at 2-3 leaf stage. Total input costs for seed, fertiliser, in-crop herbicides and foliar fungicide = $430/ha and $556/ha (Exp’s 1 and 2).(n.b. ^ɸaLabel recommends only pre-emergent use; ^ɸbLabel recommends 250-350 ml/ha)

The two sowing times in 2012 (Exp 1) were late April (canola and lupin-grain) and mid May (field peas BM and wheat) and one sowing time in Exp 2 (2^nd May 2013). The weed-free fallow commenced in early September and pea BM plots were sprayed in mid October (2012 and 2013) and re-sprayed within 14 days as a double knock. The lupin-grain treatment was spray topped in mid November 2012 in Exp 1 only with paraquat to sterilize grass seeds. All plants were kept weed-free during the summer fallow period between crops. Initial plots were 40m in length x 1.8 m with each treatment replicated 4 times.

Year 2 treatments (wheat or second break crop) imposed, input/risk categories and input costs for Exp 1 (2013) at Eurongilly, NSW:

Each of the replicated plots in all the break crop and cereal treatments from year 1 were split into three 13.3 m sub-plots, in a split plot design. Four treatments were sown in early May 2013 (Exp 1), with treatments being canola, wheat (H or L) and cereal hay (wheat). Wheat (H and L) were sown into all year 1 treatments and canola was sown into pulse, wheat or fallow year 1 treatments only. Wheat (Hay) was sown into canola year 1 treatment to act as a double break instead of sowing canola in year 2 after canola year 1. The herbicides used in all Exp’s 1 and 2 in year 2 were similar to those used in year 1 for the respective crop and input category. A significant proportion of the input costs in year 2 treatments related to the amount of urea applied. The aim was to apply nitrogen to all year 2 treatments to achieve a wheat grain yield of 7 t/ha in the high wheat treatments, 4 t/ha in the low wheat treatments and 3.5 t/ha in the canola treatments. Urea was applied at rates between nil and 330 kg/ha.Year 2 experimental details are outlined below.

1. Canola 2013 following pulses 2012 - cv Hyola® 575CL seed dressed with Jockey Stayer and Gaucho, banded with 75 kg/ha MAP (7.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha) and 100 kg/ha urea (46 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and insecticides = $352/ha.
2. Canola 2013 following wheat (H or L) 2012 - cv Hyola® 575CL seed dressed with Jockey Stayer and Gaucho, banded with 75 kg/ha MAP (7.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha) and 200 kg/ha urea (92 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and insecticides = $417/ha.
3. Wheat (Hay) 2013 following canola (H or L) 2012 - cv Gauntlet, target density 150 plants/m2; banded with 25 kg/ha MAP (2.5 kg N/ha) with Impact Endure and 40 kg/ha urea (18kg N /ha) and in-crop topdressing of 100 kg/ha ammonium sulphate (21 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and foliar fungicide = $157/ha.
4. Wheat (L) 2013 following wheat (H), pea BM or fallow year 1 - cv Gauntlet, target density 75 plants/m2; banded with 25 kg/ha MAP (2.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and foliar fungicide = $133/ha.
5. Wheat (L) 2013 following lupin (grain) or canola year 1 - Topdressed with urea at GS30 in Exp 2 only @ 50 kg/ha. Total fertiliser elements as N:P:S (kg/ha) were 22.7, 5.5, 24.4 (Exp 1). Total input costs = $133/ha (exp 1) .
6. Wheat (L) 2013 following wheat low year 1 - Total fertiliser elements as N:P:S (kg/ha) were 22.7, 5.5, 24.4 (Exp 1). Total input costs = $133/ha (Exp 1).
7. Wheat (H) 2013 following pulses 2012 - cv Gauntlet, target density 150 plants/m2; banded with 75 kg/ha MAP (7.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha) and 100 kg/ha urea (46 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and foliar fungicide = $412/ha.
8. Wheat (H) 2013 following wheat (H and L) 2012 - cv Gauntlet, target density 150 plants/m2; banded with 75 kg/ha MAP (7.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha) and 200 kg/ha urea (92 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and foliar fungicide = $478/ha.
9. Wheat (H) 2013 following canola (H and L) 2012 - cv Gauntlet, target density 150 plants/m2; banded with 75 kg/ha MAP (7.5 kg N/ha) with Impact Endure and in-crop top-dressing of 100 kg/ha ammonium sulphate (21 kg N/ha) and 260 kg/ha urea (92 kg N/ha). Total input costs including initial knock-down, in-crop herbicides and foliar fungicide = $517/ha.

Year 3 Experiment 1, at Eurongilly (2014) - Second wheat after break crops

All plots were sown to wheat cv Suntop, at a target density of 150 plants/m2, seed dressed with Dividend and fertilized with MAP dressed with Impact Endure banded with the seed at 75 kg/ha (7.5 kg N/ha). All plots had an initial knockdown of Weedmaster Argo (1.9 L/ha) and Hammer (45 ml/ha), pre-emergent herbicides of Sakura 850WG (118 g/ha) and Avadex Xtra (2 L/ha) and pre-emergent insecticide of Lorsban @ 900ml/ha. In crop insecticide was Pirimor WG (300g/ha). The difference in total input costs between the 2014 treatments relates to the amount of urea applied. The aim was to apply nitrogen to achieve a wheat grain yield of 5 t/ha for all treatments. Urea was applied at rates between 87 kg/ha and 187 kg/ha. Total urea applied and input costs for each group of treatments outlined in Table 3.

Table 3: Quantities of urea (kg/ha) and input costs from Experiment 1 year 3 (2014) for each group of sequences with crop and input types from year 1 and 2 at Eurongilly, NSW.

Results Part 1: Impact of break crops on profit and nitrogen at Junee Reefs

Junee Reefs (2011), year 1 – Break crop year

Growing season rainfall (GSR – April to October) was 216 mm which was considerably lower than the long-term average GSR of 311 mm, but heavy rainfall in February 2011 (226 mm) resulted in an annual total of 639 mm which was approximately 130 mm wetter than the long-term average (506 mm). The soil moisture profile at the beginning of the growing season was full which contributed to respectable grain yields for the various break crops and cereals grown at Junee Reefs (Table 4).

Canola yields were 3.2-3.3 t/ha with an oil content on 46-49%. The N fertiliser applied to wheat (low input = 48 kg N/ha and high input = 100 kg N/ha) had little effect on yield (4.8 cf 5.2 t/ha; respectively), but did have a large impact on grain protein (low input 10.4% and high input 12.5%). Low grain prices received for wheat in 2011 ($155/t for ASW and $203/t for AH2) resulted in canola and lentil being more profitable than both wheat and barley, and lupin to be more profitable than the low input wheat treatment (Table 4).

Gross margins achieved on-farm for wheat in 2011 may have been lower than shown in Table 4 due to downgraded wheat resulting from 150 mm rain throughout November and December 2011. The favourable gross margins received for canola and lentil relative to cereals were reflected in profit : input cost ratios of between 1.6:1 and 3:1 compared to 0.9-1.3 for wheat (Table 4). In other words, $1.60-$3 profit was recieved for every $1 spent when growing these particular break crops. This compared to < $1.30 profit for every $1 invested on wheat. No income was generated by the brown manure treatments in 2011, but they have the potential to contribute in future years through additional N supply and/or carry-over of residual soil water.

Table 4: Comparisons of grain yield, income, variable costs and gross margins of cereals and various break crops grown for grain or brown manure (BM) from year 1 at Junee Reefs, NSW in 2011. Crops arranged in order of descending gross margin.

^aNote: Grain prices used in the calculations were current at the time of harvest and assumed delivery to Junee except RR canola to Stockinbingal (extra freight cost = $5/t) and lentils to Victoria (extra freight cost = $53/t).

Comparisons of the amounts of N removed in grain at harvest in 2011 with inputs of N via applications of N fertiliser to canola and cereals, or calculated to be contributed in N fixation by legumes, and the resulting estimates of the net N balance at grain harvest are presented in Table 5. These data indicated that canola and wheat cropping represented the greatest losses of N from the system while the largest net returns of N were derived from the brown manured legume treatments and lupin grain crop.

Table 5: Measures of N removed from grain, N inputs from crops and fertilisers, net N-balance and concentrations of soil mineral N (0-160cm) prior to sowing wheat in 2012 following cereals and various break crops grown for grain or brown manure (BM) from year 1 at Junee Reefs, NSW in 2011. Crops are arranged in order of descending net N balance.

^a Includes inputs of fixed N by legumes adjusted to include N in the nodulated roots (Unkovich et al 2010), and inputs of fertilzer N to wheat, barley and canola.
^b Net N balance = (N inputs) – (N removed in grain)

Junee Reefs (2012), year 2 - First wheat after break crops

Despite 115 mm of rainfall in March 2012, residual effects of some of the 2011 treatments on soil water reserves were still measured when wheat was sown in May 2012. The stand out break crop treatments were chickpea and the brown-manured field peas and lupin which had between 40-65 mm more soil water (0-160 mm) than where wheat had been grown in 2011 (data not shown).

The concentrations of soil mineral N measured in autumn 2012 were all significantly higher after the 2011 legume treatments than following canola or cereals (Table 5). Trends in soil mineral N were related to, but were not exactly aligned with, net N balances determined in 2011 (Table 5). The enhanced availability of soil N after chickpea may have been related to the high proportion of chickpea N (6%N) partitioned below-ground in nodules, conducive to more rapid decomposition and mineralisation.

Given the 1-2 decile GSR experienced at Junee Reefs in 2012, wheat grain yields following the pulse treatments in 2011 were significantly higher than after cereals or canola in the low input wheat (Table 6). However, the application of the higher N fertiliser rates in the high input wheat in 2012 following the 2011 pulse treatments resulted in high grain protein contents, high screenings (>5%) and a reduction in grain yield suggesting that the wheat ‘hayed-off’ (Table 6).

The gross margins and profit : cost ratios calculated for the low input wheat in 2012 were greater than that achieved by the high input wheat for the same 2011 pre-crop treatments reflecting the lower production costs in a dry cropping year (Table 6).

Table 6: Comparisons of grain yields and protein content, income, gross margins and profit/cost ratios for wheat grown with low and high input 2012 and total variable costs are $315/ha and $556/ha respectively, from year 2 at Junee Reefs, NSW in 2012 following cereals and various break crops in 2011.Crop 2011 pre-treatments are arranged in order of descending 2012 gross margin.

Total production costs ranged between $72-$93/t (low input 2012) and $144-$152/t of grain produced.
* Indicates grain screenings > 5% reducing quality.

Junee Reefs (2013), year 3 - Second wheat after break crops

Annual rainfall and GSR at Junee Reefs for 2013 were 432 mm and 289 mm respectively, well below average. In April 2013, all low input 2012 treatments and the lupins BM (2012 high input) treatments were soil sampled and analysed for mineral N concentrations. Fertiliser N requirements were calculated for each low input 2012 treatment to achieve a 5 t/ha grain target yield (Data not shown).

There were no significant differences in grain yield between treatments in November 2013 with an average wheat yield of 4.46t/ha and protein concentrations ranging from 11.5-12.5% (data not shown). The most limiting factor effecting grain yield in 2013 was soil moisture. The top seven gross margins from 2013 were wheat following either a pulse or canola in 2011, with the top nine profit/cost ratios being pulse crops in 2011. The best profit/cost ratio treatments returned $1.50 to $2.00 for each $1 spent, indicating that in dry years, risk is reduced following a pulse.

Economic analyses of the sequence years 1-3 at Junee Reefs, (2011-2013)

Calculations across the three years of the experiment indicated that the average annual gross margin fell into four distinct groups (Table 7). The sequences with the highest average annual gross margins (AAGM >$600/ha per year) involved break crops in 2011. These eight treatments were the same treatments that had AAGM >$600/ha/year from the two year gross margins calculated at the end of 2012. The gross margin for canola in year 1 of phase 1 had a major positive impact on the profitability of the sequence. These break crop-wheat sequences generally had higher profit : cost ratios than the second cohort (AAGM $500-$600/ha per year) which was dominated by cereal-wheat sequences. The third cohort included the lupin and chickpea (grain crops 2011) and lupin and field pea BM low 2012/13 input crops. The final group (average annual gross margins <$400/ha per year) were the brown manure lupin and field pea–wheat sequences with high 2012 inputs (Table 7). High N fertiliser usage in the dry year of 2012 on these BM treatments resulted in very low gross margins in 2012.

Table 7: Comparisons of the mean annual gross margins ($/ha/yr) calculated for different crop sequences from three years of experimental data at years 1-3 at Junee Reefs, NSW in 2011 to 2013. Crop sequences are arranged in order of descending average annual gross margin.

^a Derived from Table 1. ^b Derived from Table 3. ^c Derived from Table 5.

Results Part 2: Impact of break crops on weeds and profit at Eurongilly

Eurongilly - Exp 1 (2012) and Exp 2 (2013), year 1 – First break crop year

Despite the GSR in 2012 (Exp 1, yr 1) at Eurongilly measuring 179 mm compared to a long-term average of 328 mm, the 2012 annual rainfall was 556 mm. Summer rainfall of 234 mm fell over February and March resulting in good soil moisture at the start of the 2012 season (Table 2). In comparision, the 2013 GSR rainfall (Exp 2, yr 1) measured 281mm with 115 mm in February-April 2013 resulting in a dry pre-season start.

a) Effect of crop sequences on ryegrass populations

The efficacy of the various research herbicide treatments used in the Eurongilly experiment are presented in Table 1. The number of ryegrass panicles measured in late spring were 1,042 and 1,804 per m² in untreated areas immediately outside the experimental plots, and ranged from 78-534 per m² under wheat (high-low) to zero under the RR-canola (high input) and bare fallow in experiments 1 and 2. No viable ryegrass seed was set at the end of year 1 (Exp 1 - 2012 or Exp 2 - 2013) by the pea BM, fallow or RR canola treatments. Cheaper, more effective ryegrass control was achieved in the break crops and fallow compared to the options available for in-crop grass management within wheat (Table 8). Pre-emergent and in-crop grass herbicides were sprayed onto the pea BM treatment to reduce ryegrass populations and DM and increase legume DM (to potentially increase amount of legume nitrogen).

The higher GSR in 2013 in experiment 2, year 1 (Table 2), resulted in an increase in ryegrass DM and panicle numbers/m² in some treatments compared to experiment 1 year 1 (i.e. wheat – low, TT canola – low, lupin – grain) (Table 8). This was very evident in the lupin - grain treatment with a 10 fold increase in panicle number and DM between the two experiments, and a doubling of ryegrass DM in the wheat – low and TTcanola – low treatments (Table 8). It was anticipated that the different levels of control would have implications for future ryegrass incidence, crop productivity and profit.

Table 8: Ryegrass control costs, ryegrass panicle numbers (m²) and ryegrass DM (t/ha) for wheat and various break crops grown for grain or brown manure (BM) and fallow in Experiments 1 and 2, year 1 at Eurongilly, NSW in 2012 and 2013, respectively. Crops arranged in order of descending ryegrass panicle numbers in experiment 1.

Note: two summer sprays of glyphosate sprayed by farmers in February 2012 and early March 2013 (cost not included)
^ KBS = killed before seedset (brown manure treatments sprayed out prior to seed set) in Exp 1, with the ryegrass counted in exp 2, but sprayed before seedset.
^a The lupin grain crop in experiment 2 was not spray topped with paraquat to sterilise ryegrass seed.
^b estimated ryegrass DM based on panicle number using regression of DM = 0.1746+0.0051x (r²=0.92)
NM – not measured. All plant material in the fallow treatment was killed before any seed set. fallow sprayed at pre sowing (April/May) and Sept (Exp 1) or July and in October (Exp 2).

Eurongilly - Exp 1 (2013) and Exp 2 (seedbank data only), year 2 – Second break or wheat crop

The season started dry in 2013 followed by wet May to September 2013 with rainfall of 259mm (Exp 1 yr 2), but overall the rainfall was below average with a GSR of 281 mm and total annual rainfall of 432mm. Only the seed bank results from Exp 2 will be presented here to complete the year 1 data set. For simplicity, no other results from year 2 (2014) or year 3 (2015) will be presented in the paper from Experiment 2. The complete set of results from Eurongilly experiment 2 will be presented.

Table 9: Ryegrass seed bank populations (m²) at the start of the experiment and after the completetion of treatments in year 1 in Experiment 1 (autumn 2013) and experiment 2 (autumn 2014) at Eurongilly, NSW.

Soil cores were taken in March 2013 (Exp 1) and March 2014 (Exp 2) from all the year 1 treatments to measure the annual ryegrass seed bank plant populations remaining in each treatment at the start of the 2^nd season. The wheat-low year 1 treatment had the highest seed bank populations, generally 10-fold more than all other treatments (Table 9). The herbicide regime of Sakura @118g/ha at sowing followed by Boxer Gold @ 2.5L/ha at 3 leaf wheat stage^ɸa and Axial @ 300ml/ha (Exp 1 only) in early July combined with the increased wheat sowing rate (150 plants/m² in high wheat cf 75 plants/m² in low) were effective at reducing ryegrass panicle numbers in the wheat-high treatment to between 30-78 ryegrass panicles/m² at the start of year 2 (Table 9) and reducting the seedbank population from 1815 and 2775 seeds/m² by 50-60% respectively, in experiment 1 and 2 (Table 9). (n.b. ^ɸaLabel indicates only pre-emergent use)

The combination of good GSR in 2013 and the failure to spraytop the ryegrass in lupin crop in experiment 2 year 1 resulted in a significant increase (75%) in ryegrass seedbank (Table 9). Similarly, the canola (L) performed poorly in experiment 2 with a 30 per cent increase in ryegrass seedbank, whereas there was a 70 per cent reduction in seedbank in experiment 1 – from 1800 to 505 plants/m² (Table 9). Generally, the greatest reduction in seedbank population was achieved by the fallow, RRcanola (H) and fieldpea BM year 1 treatments.

Table 10: Annual ryegrass panicle numbers (m²) following various cereal and break crop options in years 1 (Exp 1) followed by either wheat (high, low or hay) or canola in year 2 (Exp 1) at Eurongilly, NSW between 2012 and 2014.

By spring in year 2 (2013), there were significant differences in panicles/m² with four distinct categories (0-8, 14-71, 192-388 & >643 panicles/m²) (Table 10). The treatments with the greatest reduction in ryegrass panicle numbers and least amount of ryegrass DM by the end of year 2 were any sequence that included canola (zero panicles/m²) at the end of year 2 and < 8 panicles/m² following wheat (H) yr 2 treatments after a pulse legume crop or fallow (Table 10). The wheat-hay or wheat (H) treatment following a canola crop (that had controlled ryegrass) were also effective at reducing ryegrass numbers. The highest number of ryegrass panicles were found in the untreated areas, followed by the sequence wheat L) year 1 / wheat (L) year year 2, then by any wheat (L) treatments irrepective of panicle numbers in year 1 (Table 10).

Eurongilly - Exp 1 (2014), year 3 – Third crop (wheat in all treatments).

With all sequences sown to wheat in year 3 (sprayed with Sakura pre-em at 118g/ha), the sequences that had the biggest reduction on panicle number and ryegrass DM by November 2014 (end yr 3) were double break sequences i.e. lupin-grain/canola, peasBM/canola and fallow/canola (Table 10). The main 2013 treatment effects (year 2) continued into 2014 (year 3) with significantly less panicles in order of: canola < hay = wheat (H) < wheat (L), and year 1 effects: fallow < pulses < canola = wheat (H) < wheat (L) and interactions categorised into groups of (0-30, 60-166, 199-370, >536 panicles/m²) (Table 10). Generally, double break sequences or those where wheat (H) treatments were grown following treatments with bare soil or less stubble from year 1 had significantly fewer panicles. Where Sakura was sprayed for three consecutive years in the wheat (H) year 1 / wheat (H) year 2 wheat sequence, there was a trend for ryegrass panicle numbers to increase over the 3 years (Table 10).

By autumn 2014 (year 3), there was a further significant 2.5 fold increase in ryegrass seedbank populations from 5492 seeds/m² to 13148 seed/m² following two consecutive wheat (L) treatments (Table 10). Comparatively, seedbank numbers reduced to 124 seeds/m² where canola (H) 2012 was followed by wheat hay (2013), and double breaks involving legumes, canola, fallow or hay resulted in the lowest seed banks following the three year sequences (Table 10). Main effects from year 1 (2012) and year 2 (2013) treatments were still apparent after the conclusion of the experiment in March 2015 (start year 4), with the year 2 treatments having a greater effect with significantly higher seedbank numbers remaining in order of: wheat (L) > wheat (H) > wheat (hay) > canola (meaned data not shown). The double break of fallow or lupin followed by canola reduced the seedbank to the lowest number (Table 10). The expensive herbicide costs ($142/ha) associated with consecutive wheat (H) treatments resulted in a significant reduction in seed bank by November 2014 (366 plants/m²), but was not as effective as sequences involving break crops (Table 10).

b) Effect of crop sequences on crop yields and gross margins

In year 1 the most profitable crops were canola (H) and (L) which returned gross margins of $1259 and $1166 /ha, respectively. The next most profitable crops were lupins (H) @ $683/ha, wheat (H) @ $257/ha, wheat (L) @ $250/ha, with the brown manure or fallow treatments resulted in negative gross margins (-$45 to -$250/ha). The cool end to the season and the early November rain in 2012 (Exp 1 year 1) assisted canola to yield similarly to wheat (farmer header yields were close to 3t/ha for the region). Canola and lupin grain were more profitable than both wheat treatments with a profit : cost ratio of between 1.8 and 2.6 compared to 0.4 to 0.9 for wheat (Table 11). The wheat (L & H) treatments performed poorly even though between 50 and 100 kgN/ha were applied to the treatments, respectively (starting soil mineral N was 87 kgN/ha 0-150 cm). However, competition with ryegrass may have reduced yield (Table 9 - ryegrass DM 1.6 t/ha and 0.25t/ha in wheat low and high). The canola price was high 2012 ($490/t) which assisted the high canola GM.

Table 11: Comparisons of grain yield, income, variable costs and gross margins of wheat and various break crops grown for grain or brown manure (BM) or fallow from Experiment 1, Year 1 at Eurongilly, NSW in 2012. Crops arranged in order of descending gross margin.

^aNote: Grain prices used in the calculations were current at the around the time of harvest and assumed delivery to Junee except RR canola to Stockinbingal (extra freight cost = $5/t).

() brackets indicate protein concentration as a percentage in wheat grain.

In year 2, the treatments with the highest gross margin were canola following fallow or brown manure treatments (> $1000/ha) with canola following wheat (H) or lupins (H) returning ~$900/ha (Table 12). Wheat yield and grain protein were significantly higher in the wheat (H) year 2 treatments than the wheat (L) year 2 treatments in 2013 with an average grain yield and protein content of 4.9 t/ha and 12.7% cf 3.1 t/ha and 12.2% (Data not shown). Grain yields and profit were generally highest for canola or wheat (H) following a pulse or fallow, with the lowest yields and gross margins returned from any combination that had wheat(L) in either 2012 or 2013, except following fallow (Table 12). The wheat (Hay) treatments following canola yr 1 (2012) resuted in reasonable yields (7.4-8.1 t/ha DM) and gross margins of near, or above $600/ha (Table 12).

Table 12: Comparisons of grain yield and annual gross margins in years 1 to 3 (2012 to 2014), average annual 3 year gross margin and average profit/cost ratio over the 3 years for Experiment 1 at Eurongilly, NSW following cereals and various break crops in 2012. Crop 2012 pre-treatments are arranged in order of descending average 3 year gross margin.

Over the three years, the most profitable 3 year sequence was canola (H) - wheat (H) – wheat, with an average GM of $883/ha/yr. Sequences with the highest average annual gross margins >$800/ha/yr where treatments that had canola in year 1, with the next most profitable group having lupins (H) in year 1 or canola year 2 (> $600/ha/yr). The third group (>$500/ha GM) included sequences of fallow, combinations of wheat (H or L), with the final group involving sequences with BM crops followed by wheat (H) or wheat (L) (Table 12). The sequences with the lowest risk and a profit of $2 for each $1 spent (highest profit/cost ratio) were wheat (L) following either a canola (L), lupin or fallow treatment in year 1.

Interaction between crop treatments and ryegrass plant populations

The effect of the high and low input treatments on ryegrass control and ultimately wheat grain yield can be seen in Figure 1. The high input treatment (open symbols) significantly reduced ryegrass DM and increased wheat grain yield. This compares to the increase in ryegrass DM under the low input treatments (closed symbols) resulting in a reduction in wheat grain yield of 450 kg/ha for every 1 t/ha of ryegrass DM (Figure 1).

Figure 1: Relationships between ryegrass dry matter (DM) and wheat grain yield following high and low input treatments in wheat in 2012 and 2013 (Exp 1 and 2) at Eurongilly, NSW.

Table 13: Average annual gross margin over three years compared to ryegrass seedbank in March 2015 in Experiment 1 at Eurongilly, NSW. Crop 2012 pre-treatments are arranged in order of descending average 3 year gross margin. Grey strips indicate double break sequences.

It is evident that the combination of a fallow or break crop in year 1 followed by a second break crop in year 2 resulted in the greatest reduction in ryegrass seedbank populations and panicles after three years of sequences (Table 13), and can still be profitable. Where herbicide resistent ryegrass is a major problem, an alternate sequence to reduce seedbank to extremely low levels and be profitable could be - Year 1: Wheat-hay (sprayed afterwards), Year 2: pulse-grain (spraytopped), Year 3 RRcanola.

Conclusions

Break crop effects on income and nitrogen

Results from experimentation undertaken in southern NSW between 2011 and 2013 have shown that canola and legume break crops can frequently be as profitable, and in a number of instances considerably more profitable, than wheat. Canola was consistently the most profitable break crop. However, legumes provide additional rotational benefits for subsequent crops by increasing soil N supply.

Wheat following break crops was consistently more profitable than wheat on wheat. Sequences with canola were largely profitable due to the high returns from canola itself. Crop sequences involving legumes can be profitable due to increased wheat yields and lower costs of production.

Long fallows and pulses grown for brown manure (BM) lose money in the year they are grown, but achieve excellent grass weed control, high N inputs and potential residual carry-over of soil water. The use of fallow or BM enhances the yield of following wheat crops, but the increased grain yields were insufficient to fully compensate for the loss of income in the first two years. The net result was the 3 year gross margin using BM was the least profitable of any break crop treatment, and the high available soil N following BM can also increase the risk of haying-off of a following wheat crop in an exceptionally dry year if additional fertiliser N is applied.

It was demonstrated that cheaper, more effective ryegrass control could be achieved by many of the alternative break crops compared to the options available for best in-crop grass management within wheat where there is a high level of herbicide-tolerant ryegrass. It was apparent that there is a requirement for at two break crop years to control herbicide resistant ryegrass with the most profitable options being RRcanola followed by a wheat (H) treatment. However, after the three year sequence, this treatment still had significantly higher seed bank numbers than other less profitable options. The lupin crop grown for grain followed by the RRcanola, or RR-canola followed by wheat (Hay) may be more effective in terms of gross margins in the long run and reducing seed bank populations. Sequences involving fallows and brown manures are very effective at controlling ARG, and reduced production risk in subsequent years, but were not as profitable as continuous cropping. A pulseBM crop followed by RRcanola is most likely to be beneficial where annual ryegrass resistance and low soil N are problematic.

Acknowledgements

We are grateful to the Grains Research and Development Corporation (GRDC project CSP000146) for financial support to undertake this research. We are also indebted to our farmer collaborators, Rob and Bernard Hart from Junee Reefs, Tom, Simon and Angus Brabin and John Martin from Eurongilly for allowing us access to their land to establish on-farm experimental trials.

References cited

Broster JC, Koetz, EA, Wu H (2011) Herbicide resistance levels in annual ryegrass (Lolium rigidum Gaud.) in southern New South Wales. Plant Protection Quarterly 26(1) 22-28.

Isbell (1996) ‘The Australian soil classification.’ (CSIRO Publishing: Melbourne)

Preston C, Boutsalis P, Malone J, Abu-Yeboah Patricia, Kleemann S, Saini RK, Gill Gurjeet (2013) Maintaining the best options with herbicides. 2013 NSW GRDC Grains Research Update for Advisers, Temora.

Contact details

Tony Swan
CSIRO Plant Industry, GPO Box 1600 Canberra ACT 2601
tony.swan@csiro.au