1. A comparison under grazing of pasture production, pasture N content and soil mineral N levels between granular urea and ONEsystem ® on two contrasting dairy farms in New Zealand Bert Quin, Allan Gillingham, David Baird, Stewart Spilsbury and Maurice Gray Presented to the 2015 New Zealand Grassland Society Conference, Masterton, New Zealand, 2-5 November 2015

2. Why develop ONEsystem®? A brief history of fertiliser urea 1.The Haber-Bosch process for producing urea was developed in 1922. Urea is relatively cheap to produce per kg N, but agronomically inefficient, especially when surface applied. 2.Many attempts to improve urea have been made, most centered around slowing the release of the urea from the granule by coating it with am impervious membrane such as polymers and/or elemental S. Wikipedia lists about 50 currently variations on sale. Edmeades (2015) contains no new information. All coated urea products ‘work’ to a degree, but high manufacturing costs have and will conitinue to restrict their use to high-value crops and recreational turf application.

3. 3. They are particularly unlikely to have a role under grazing, because (i) dairy farmers want their N response now, not in x months, and (ii) hoof pressure easily destroys the efficacy of the coating. Why develop ONEsystem®? A brief history of fertiliser urea 4. Granular urea with nbpt (SustaiN) was far more efficient, and much, much cheaper. This is why it is already 40% of the NZ market. 5. However, while an improvement, SustaiN’s average EDM factor (extra kg DM/kg N applied) is still only 13, cf 10 for regular urea. 6. With increasing availability internationally of prilled (0.8 – 2.8 mm) urea, it was logical to investigate its optimisation for use on pasture. At 30 kg N/ha, 450 prills are applied per square meter (a close fit with pasture plant density) compared to only 45 for granular urea.

4. The ONEsystem ® trials  The trials were laid down, monitored and measured by experienced, independent scientist Dr Allan Gillingham and field technician Maurice Gray.  All trial data were analysed by Dr David Baird, an international independent biometrician of high standing.  All 3 trials were conducted under grazing. The 2 main trials, one each in mid-Canterbury (irrigated) and Waikato (rain-fed), ran for 4 grazing rotations from early-spring to early summer, with 4 applications of fertiliser N.

5. The ONEsystem ® trials  A second trial in mid-Canterbury was designed to investigate the relative importance of urea particle size, water spray and nbpt. This trial ran for one grazing rotation only, in early summer.  The conduct of the trials under grazing was considered crucial, despite the increased variability in pasture production, to produce ‘real-life’ results. Small-plot fertiliser N trials are extremely susceptible to the generation of misleading results in trials lasting more than 2-3 months, because of the artificial depletion of soil N fertility in the control plots. This is particularly the case if 70-80% clippings return is not practised.

6. Trial treatments  Mid-Canterbury Trial 1 and Waikato (Trial 2) Control plus 3 rates of N of each fertiliser N product (same for each product). 4 applications were made, immediately after grazing to 1200-1500 kg DM/ha.  Mid-Canterbury Trial 3 A single rate (28 kg N/ha) of both granular and prilled urea were applied, both applied (i) in dry form, (ii) with water spray added (50l/ha), and (iii) with a solution of nbpt spray added (50 l/ha water, nbpt application 30 gm/ha). This trial ran for 8 weeks following a single fertiliser application, to ensure the full N response was measured.

7. Reece Agrispred 460N tow-behind spreader (with tanks fitted for application of nbpt spray)

8. Soil data, herbage composition and N treatments Table 1 Soil characteristics, herbage composition and N treatments at Canterbury (Site C) and Waikato (Site W) experiments. ------------------------------------------------------------------------------------------------------------------- soil ‘Quick’ test pasture sp (%) +N rates (kg N/ha per application) pH P K S CEC rye clvr poa wds granular urea ONEsystem® ______________ ________________ ____________________________ Site C 5.6 26 4 11 11 99 0.5 0.1 0.5 14.1, 28.2, 42.3 14.0, 28.0, 43.0 Site W 6.6 42 12 11 32* 68 7 22 3 27.4, 54.8, 82.2 18.3, 36.6, 54.9 _____________________________________________________________________________ *note high effective CEC at Site W due to pH

9. Dry matter responses for individual periods, trials 1 and 2 Figure 1 Pasture extra dry matter (EDM) at Canterbury and Waikato with increasing rates of N as granular urea and ONEsystem® on four occasions. The LSD at the 5% level between means is shown.

10. Overall DM results, trials 1 and 2 74 17 126 10 120 10 50 24 Figure 2Total pasture extra dry matter (EDM) at Canterbury and Waikato from increasing rates of N applied as either granular urea (dashed lines) or as ONEsystem® (solid lines). The dotted vertical lines with show the predicted total application of N for granular urea and ONEsystem® to achieve EDM of 1250 kg/ha at either site. The numbers circled are the respective EDM factors (kg extra DM/ kg N applied). The least significant difference at the 5% level between means is shown.

11. Total N uptake, trials 1 and 2 Average plant nitrogen uptake (kg N/ha) over 4 periods prior to grazing at the Canterbury and Waikato sites, with increasing rates of N applied as either granules or ONEsystem®. The least significant difference at the 5% level between means is shown.

12. Soil nitrate-N, trials 1 and 2 Soil nitrate-N levels (ppm) averaged over 4 applications at the Canterbury and Waikato experiments sites with increasing rates of N applied as either granules or ONEsystem®. The least significant difference at the 5% level between means is shown.

13. Trial 3 – Effect of particle size, water and nbpt ___________________________________________________________________________ Control (nil N) dry fertiliser N N +water N + water + nbpt Granular urea 1727 a 2016 ab 1780 ab 2121 bc (9.6) (1.8) (13.1) Prilled urea 1727 a 2344 cd 2579 d 2432 d (20.6) (28.4) (23.5) ___________________________________________________________________________ LSD 5%: 297. Significance : urea *** ; urea form (granular/prilled) *** ; urea form x method * Duncan’s Multiple Range Test: Treatments results that are not followed by the same letter differ (P<0.05) from one another. Table 4 Canterbury Site C2. Pasture yield (kg DM/ha) obtained with a nil N control and single applications of N (28 kg N/ha) of dry, wetted and wetted+nbpt versions of granular and prilled urea. EDM factors are in brackets.

14. Conclusions * On the mid-Canterbury Trial 1, ONEsystem® increased DM response by 240% compared to granular N. At typical rates of application of N (ca. 30 kg N/ha) the EDM factors (extra dry matter per kg N applied) were 24 for ONEsystem®, compared to 10 for granular urea (P<0.01). * On the Waikato Trial 2, ONEsystem® increased DM response by 170% compared to granular N (P<0.1). At typical rates of application of N (ca. 30 kg N/ha) the EDM factors were 17 for ONEsystem® compared to 10 for granular urea(P<0.1).

15. Conclusions *To maintain the advantage to ONEsystem® in highly organic soils in warmer temperatures, either the rate of nbpt application needs to be doubled, or the nbpt needs to be applied directly to the prills. * In Trial 3, the biggest effect was the increased response achieved using prills instead of granules. Dry prilled urea produced significantly more EDM than dry granules. Water and nbpt contributed. (Prills+water) produced higher EDM than (granules+water), and both (prills+water) and (prills+water+nbpt) produced more EDM than any granular form. (Prills+water+nbpt), ie ONEsystem®, produced an EDM factor of 23.5, compared to 13 for (granules+water+nbpt), ie wetted SustaiN, and 10 for dry granular urea.

16. Thank you for your attention