Previous PageTable Of ContentsNext Page

Nitrogen stress during tillering decreases lodging risk and increases yield of irrigated bread-wheat (Triticum aestivum) in north-eastern Australia.

Allan Peake1, Kerry Bell2, Nick Poole3 and John Lawrence1

1 CSIRO Sustainable Agriculture Flagship. Email: allan.peake@csiro.au
2
Queensland Department of Agriculture, Forestry and Fisheries
3
Foundation for Arable Research, NZ.

Abstract

Irrigated cotton growers in north-eastern Australia have recently grown record areas of irrigated wheat (notably in 2008) in response to increased grain-prices, but wheat yields in the region have been severely constrained by lodging. Irrigated experiments were conducted at Gatton in 2009 and 2011 to assess the ability of canopy management techniques of delayed nitrogen (N) application and low plant populations to decrease lodging risk in the northern region. High leaf area index at the end of tillering was associated with increased lodging. Maximum yields for irrigated experiments were generally achieved when soil + fertiliser N at sowing was less than 100 kg/ha, with low N treatments having less lodging and yield increases of up to 1 t/ha-1 over high N treatments. Increasing plant density above 100 plants/m2 increased lodging and decreased yield in high N treatments. The highest yielding treatments had the least lodging, a harvest index of 0.45, and <450 tillers/m2 at the end of tillering. We conclude that canopy management techniques can be used to increase yields and decrease lodging in irrigated wheat in the northern region, but are different to techniques used for irrigated wheat growing in southern Australia. The responses observed may have been reliant on irrigation during tillering to ensure low N levels were fully available to the crop. Further study is needed to determine the importance of early-season irrigation in maintaining yield on low N paddocks.

Key Words

Wheat, nitrogen stress, lodging, yield, irrigation, canopy management

Introduction

In 2008, irrigated growers across north-eastern Australia planted record areas of irrigated wheat in response to high grain prices and good water availability. Although many growers obtained agronomic recommendations for irrigated wheat growing from southern Australia (e.g. Lacy and Giblin, 2006), large areas of the crop lodged and experienced yield losses partially due to shattering, and sprouting from wet weather during harvest. Fields expected to achieve yields of 7-8 t/ha only yielded 3-4 t/ha (Peake and Angus, 2009), and farm-gate losses were conservatively estimated at $20 million through lost yield and increased harvest costs, although the actual figure may have been much greater.

Susceptibility to lodging is linked to tall, dense canopies which are more likely to develop due to high soil or fertiliser nitrogen (N) levels, early sowing, long-season varieties and high seeding rates (Berry et al. 2004). Research in the UK in winter wheat has demonstrated that the shading effect of a thick canopy weakens stems and surface roots (Sparkes and King, 2008), while the increased height and biomass of a high-input crop produces taller, heavier plants that are more likely to bend the stem or displace the roots.

While plant breeding could be used to improve lodging resistance of local germplasm, it is not a short term solution. Additionally, the experience of high-yield winter wheat production in the UK and New Zealand shows that specific agronomic practices should be used to manage the crop canopy for maximum reduction in lodging risk even when lodging resistant germplasm is available. Therefore the aim of this paper was to determine if the canopy management techniques of ‘in-crop N application’ and reduced plant population could be used to reduce lodging risk in irrigated spring wheat in the northern grain production region of Australia, while still achieving high yields.

Methods

Field Trials

Trials were conducted at the Gatton CSIRO research station (S27° 32.470’ E152° 20.2') in 2009 and 2011 to investigate in-season nitrogen (N) fertiliser application in combination with different plant populations across two varieties; the long season variety EGA Gregorymall&longPBR.png and the quick maturing variety Kennedymall&longPBR.png. Trials were fully irrigated with overhead spray lines, and standard plots were 7 rows (22.5cm row spacing) separated by 60cm wide wheel-track gaps. An alternate plot/row configuration was also sown in 2009 to investigate yield potential and light interception of 1m wide vs 2m wide irrigation beds. These plots contained 2 sets of 3 rows of wheat separated by a 60cm gap, representing 2 x 1m irrigation beds, covering the same area as the standard plots which represented 2m wide irrigation beds. All plots were sown 7m long, but slashed to be 5m long on the day of harvest.

In 2009, four adjacent experiments were conducted under the same irrigation system. The basis of each individual experiment was a four treatment factorial containing two seeding rates (low = 100 seeds/m2; high = 200 seeds m2) and two bed configurations (1m vs 2m), with a given bed configuration used exclusively in each planting run to create a split plot design. Two further treatments were added in addition to the factorial design where 50 kg/ha of urea (applied by hand and watered in) was added shortly after sowing in combination with the high seeding rate for both bed types. Two of these experiments (one for each variety) were conducted on low soil nitrogen (on a strip where a forage sorghum crop had been recently grown and baled specifically to reduce soil N levels) and the other two (one for each variety) were conducted on soil (fallowed) with a higher nitrogen level. Paddock history of the two areas was identical prior to the forage crop, and soil tests did not reveal any nutrient deficiencies (other than N) in either paddock history (see Table 1 for residual soil N levels). The varieties were sown on different sowing dates (EGA Gregorymall&longPBR.png = 13th May, Kennedymall&longPBR.png = 5th June) to optimise yield potential for the phenology of each variety.

In 2011 all treatments were included within the same split-plot experiment and sown exclusively with the standard 2 m wide plots. A forage sorghum crop was grown and baled in the summer of 2010/2011 to ensure low levels of residual soil N for the entire experimental area. The factorial design included Kennedymall&longPBR.png and EGA Gregorymall&longPBR.png, at three N regimes and two seeding rates. The varieties were sown in separate ‘split-plots’ within each rep on 13th May (EGA Gregorymall&longPBR.png) and the 3rd of June (Kennedymall&longPBR.png). Soil residual N and N application timing for both 2009 and 2011 experiments is included in Table 1.

Table 1. Soil and applied nitrogen regimes for the different treatments in 2009 and 2011.

 

2009

 

2011

Low N

Low N +50

High N

High N +50

 

Low N

Medium N

High N

   

kg/ha N

     

kg/ha N

 

Sowing Soil N (to 180 cm)

60

60

125

125

 

15

15

15

N treatment

               

Fertiliser N (sowing)

 

50

 

50

 

0

50

150

Fertiliser N (GS31-32)

190

140

125

75

 

200

150

50

Fertiliser N (Flag Leaf)

50

50

50

50

 

50

50

50

Total Soil + Fertiliser N

300

300

300

300

 

265

265

265

Lodging Ratings and Statistical Analysis

Lodging ratings were taken regularly to track the progression of lodging over time. The lodging rating for a given day was the average stem angle from vertical for the whole plot. Average lodging during grainfill (AvLodgeGF, %) was calculated as the average lodging rating between anthesis and harvest. Yield was calculated at 12% moisture. Experiments were analysed using spatial analysis techniques in Genstat. In 2009 the four experiments were analysed together in a multi-environment/variety analysis (with a split-plot design within each of the four experiments due to the 1 m vs 2 m bed comparison), to determine if agronomic treatments interacted with variety or residual soil N level. Two analyses were conducted: the first analysed the six treatments across site/variety combinations, the second examined only the factorial design of four bed-type × seed rate combinations. Square root transformation was necessary for AvLodgeGF prior to analysis, and the results reported have been back-transformed. The analysis in 2011 investigated the factorial combination of seed rate, N-regime and variety.

Results

The low nitrogen regimes induced nitrogen stress in both 2009 and 2011. In 2009, the low N plots were moderately yellowed by GS32 and had a distinctly different canopy structure (shorter and more upright leaves) than the high N plots. In 2011, the low N plots showed extreme yellowing and stunting by the end of tillering, and the decision was taken to apply the in-season fertiliser at an earlier growth stage (GS31). Final plant density varied between years and treatments, with Kennedymall&longPBR.png averaging 65% establishment in both years, while EGA Gregorymall&longPBR.png had worse establishment than Kennedymall&longPBR.png in 2009 (40%) but better establishment (85%) in 2011. The different agronomic treatments created a large range in leaf area index (LAI) at the end of the vegetative phase, which was closely related to average lodging during grain filling (AvLodgeGF) in both varieties in 2009 (Figure 1a) and 2011 (data not shown).

small&longPBR.png
small&longPBR.png
small&longPBR.png
small&longPBR.png
small&longPBR.png


Figure 1: (a) Average grainfill lodging vs LAI at GS31 and (b) average lodging and grain yield for the four 2009 experiments. All means for grain yield and lodging in (b) are significantly different (p<0.05).

In 2009, the experiment-mean yield and AvLodgeGF was significantly different (p<0.001) between each experiment. Increased yield was associated with decreased lodging, with the highest experiment mean yield found in the least lodged experiment (Figure 1b). Plant population interacted with experiment for grain yield (p<0.05). The low plant density had significantly higher yield than the high plant density by 0.5 t/ha in the ‘high N’ Kennedymall&longPBR.png experiment, but no significant differences in grain yield was observed between plant densities in the other experiments. AvLodgeGF was numerically greater in the higher plant density for all experiments, but no significant differences were observed. Significant differences were observed in the six treatment analysis for both yield and AvLodgeGF (p<0.01) which showed no interaction with experiment (p>0.05). On average across experiments, the +50N treatment significantly (p<0.05) increased AvLodgeGF in the 2m bed treatment (51%) in comparison to the low N treatments (30%), and had a near-significant increase (p=0.09) in the 1 m beds, but did not significantly decrease yield in relation to the standard sowing N treatments in either the 1 m or 2 m beds. On average across experiments, 2 m beds were significantly (p<0.05) higher yielding (6.1 vs 5.7 t/ha) and more lodging susceptible (AvLodgeGF =28% vs 22%) than 1m beds.

In 2011, the main effect of plant density was significant (p<0.05) for both yield and AvLodgeGF, with the lower plant density having lower AvLodgeGF (13% vs 18%) and higher yield (6.4 vs 6.0 t/ha) than the higher plant density on average across all N rates and both varieties. The only significant higher order interaction for either yield or AvLodgeGF in 2011 was the variety × N regime interaction, which was significant (p<0.05) for AvLodgeGF, and approaching significance (p=0.07) for yield. The low N and medium N treatments in Kennedymall&longPBR.png were not significantly different for either yield or AvLodgeGF, but did have significantly (p<0.05) more lodging and lower yield in the high N treatment (Figure 2a). In the EGA Gregorymall&longPBR.png treatments, no significant differences were observed among N regimes for AvLodgeGF (Figure 2b), but grain yield was significantly lower in the low N treatment.

The highest yielding treatment in 2009 was the low seeding rate, 2m bed treatment in Kennedymall&longPBR.png ‘low N’ experiment. It had a yield of 7.4 t/ha at 12% moisture and a harvest index of 0.45. It had 430 tillers/m2 at GS32 (340 less than the equivalent treatment in the ‘high N’ Kennedymall&longPBR.png experiment) and 350 tillers /m2 at maturity. In 2011, the highest yielding treatment was the low seed rate, medium N treatment in Kennedymall&longPBR.png. It had a yield of 7.3 t/ha, a harvest index of 0.45, and had just 320 tillers/m2 at GS31 (210 less than the equivalent high N treatment), and 350 tillers m2 at maturity.

Discussion

The experimental results show that agronomic treatments which reduce crop canopy size during tillering (evidenced in reduced leaf area index) reduced lodging of two spring bread-wheat varieties under high input conditions, in agreement with the experience in high-yielding winter-wheat production (Berry et al., 2004).



Figure 2: Yield and average lodging during grainfill across 3 N regimes for (a) Kennedymall&longPBR.png and (b) EGA Gregorymall&longPBR.png in 2011. Means on the same response curve with different superscript letters are significantly different (p<0.05).

Lower plant densities than the standard dryland density (<100 plants/m2) and in-crop nitrogen application generally reduced lodging and increased grain yield compared to higher plant densities, and N application at sowing. Some severely nitrogen-stressed plots were able to recover and achieve high yields when soil N was just 15 kg/ha at sowing (in 2011) and no further N was applied until GS31, but this effect was not consistent across both varieties, as severe N stress during tillering reduced yield of EGA Gregorymall&longPBR.png in 2011. The results may have been reliant on the frequent irrigation during tillering, which would have ensured that the low N levels were fully available to the crop. Further investigation is required into the optimum N timing and irrigation regime for different commercial varieties.

In 2009, the best yields were obtained in both varieties on the low soil N paddock, when residual soil N was just 60 kg N/ha. In 2011, the best yields for both varieties were achieved when 50 kg/ha N was added at sowing to 15 kg/ha of soil residual N. The optimum soil + fertiliser N at sowing under non-limiting conditions for reduced lodging risk is probably between 50-100 kg/ha N under fully irrigated conditions, given that the addition of 50 kg/ha N at sowing increased lodging in 2009. This N level also agrees with that recommended for winter wheat (Berry et al. 2004).

Lacy and Giblin (2006) recommended 100-120 kg/ha N at sowing, 150-200 plants/m2 established, and 5-800 tillers/m2 at GS30 for irrigated wheat production paddocks targeting 8 t/ha in southern NSW. The seed rates, N regimes and tiller numbers at GS31 identified herein for minimising lodging risk while maintaining yield potential were lower than those recommended for southern NSW. However the recommendations from southern Australia were utilised by northern region growers in 2008, and it is probable that incorrect application of these recommendations outside their intended environment contributed to the widespread lodging that was experienced.

Conclusion

We conclude that canopy management techniques can be used to increase yield and decrease lodging in irrigated wheat in the northern region, but are different to those recommended in southern Australia. Further study is needed to determine the role of early-season irrigation in maintaining yield on low N paddocks.

References

Berry PM, Sterling M, Spink JH, Baker CJ, Sylvester-Bradley R, Mooney SJ, Tams AR, Ennos AR (2004) Understanding and Reducing Lodging in Cereals. Advances in Agronomy 84, 217-271.

Lacy J , Giblin K (2006) Growing eight tonnes a hectare of irrigated wheat in southern NSW. Primefacts #197, NSW DPI.

Peake A.S., and Angus J.F. (2009). Increasing yield of irrigated wheat in Queensland and Northern NSW. In GRDC Northern Region Grains Research Updates, Goondiwindi, 3-4 March, 2009.

Sparkes DL and King, M (2008). Disentangling the effects of PAR and R:FR on lodging-associated characters of wheat (Triticum aestivum). Annals of Applied Biology, 152, pp. 1-9.

Previous PageTop Of PageNext Page