This webinar will be comprised of two presentations: Discovery and application of a generalised nitrogen response function for global cereals by Hans. van Grinsven, PBL Netherlands Environmental Assessment Agency; and A new method to predict grain protein in milling wheat during the growing season by Mechteld Blake-Kalff, Hill Court Farm Research, UK.
Discovery and application of a generalised nitrogen response function for global cereals
Hans J.M. van Grinsven1, Renske Hijbeek2 and Hein F.M. ten Berge2
1Department of Water, Agriculture and Food, PBL Netherlands Environmental Assessment Agency, The Netherlands.
2Wageningen Plant Research, The Netherlands.
A crucial relationship for agronomic and food system analysis is the response of crop yield (Y) to the addition of nitrogen fertiliser (Nrate). Long-term field experiments (LTEs) are essential for correctly determining the economic and environmental performance of alternative agricultural practices. However, the vast majority of N response data are derived from trials lasting only one or two years (STEs). Typically, marginal N response from STEs is about twice as low as for LTEs, implying that caution is justified when basing regional agronomic and environmental effects of changing N regimes on STEs.
In view of the generic principles governing N transformations and uptake when N input, crop yield and soil N pools are in a near steady state, we hypothesised the existence of a generic LT N response relationship for global cereals. To test this hypothesis, we scaled N response by using indexed yield (Y/Ymax) and total N availability (sum of N input from fertiliser and non-fertiliser sources – here referred to as SN), and fitted scaled N response by 2nd order polynomials with zero intercept (curve referred to as SNR).
We first derived SNRs for winter wheat between 1985 and 2018 at Rothamsted experimental station in the UK (running for >150 years: R2 0.954 for wheat in rotation and R2 0.903 for continuous wheat), and for continuous corn between 2000 and 2010 at the Southwest Research-Extension Center in Kansas, USA (running for >50 years; R2 0.934). Next we found similar SNRs (R2 0.818) for the combined results of 25 less extensive LTEs (running >15 years) for winter wheat, barley and corn in Europe, USA and Asia, with maximum yields between 2.8 to 12.8 t/ha, N rates from 0 to 300 kgN/ha and under a wide range of practice, soils and climate. We did not find LTEs for Africa, but could demonstrate applicability of our SNRs. For paddy rice we had 4 LTEs but the resulting SNR was different.
SNRs can be transformed back to unscaled N response curves for common situations where no LTEs are available, by using site specific estimates of specific SN (mainly from N deposition biological N fixation) and Ymax. The scaled, generalised N response relationship was used to construct a globally applicable function and diagram between agronomic efficiency and Ymax. This can be used to support the development of strategies to increase regional cereal sufficiency, balancing efforts to increase Nrate and Ymax.
Finally, we developed LT N response approaches, including N removal and N losses, to assess economically optimum N rates from the farm perspective, accounting for the cost of N fertiliser use, and the societal perspective by also accounting for the cost of N pollution. For this we could only use N data from the LTEs at Rothamsted. Ranges of these optimum N rates show that the safe operating space of N fertiliser application narrows when aiming at the combined ambitions of adequate farm income, regional cereal sufficiency and acceptable levels of N pollution.
A new method to predict grain protein in milling wheat during the growing season
Mechteld Blake-Kalff 1, Laurence Blake1 and Allison Grundy2
1Hill Court Farm Research, UK
2CF Fertilisers UK Ltd., UK
Growers receive a premium on their milling wheat if they reach protein levels above 12.5%. Usually the larger the yield, the lower the protein levels will be, unless extra Nitrogen is applied. It is difficult to ascertain mid-season whether more N is necessary to achieve the required protein level, because relatively small differences in N nutrition are involved; 30-40 kg of N might be enough to push protein over the threshold.
In this paper we describe the development of an accurate test to predict final protein in grain, using samples taken mid-May to mid-June, so that it is possible to add extra fertiliser if necessary. The test measures signals in the roots that are involved in the regulation of N uptake in response to the external N supply. Generally, when plenty of nitrogen is available, the N uptake efficiency will decrease and vice versa.
Measuring these regulatory signals provides a reliable assessment of the N status of the plant, and from this the final grain protein level can be predicted. Despite the difficult growing conditions in 2019, we were able to correctly predict protein levels as being < 12%, between 12 – 12.5%, and > 12.5% in 72% of samples, which all received between 180 and 360 kg/ha N fertiliser. Further field experiments have been conducted in 2019 and 2020, so that additional results can be presented in this webinar.
There will be a written paper accompanying this webinar.