The growth, development and nitrogen dynamics of white clover seed crops: A thesis submitted in partial fulfilment of the requirement for the Degree of Doctor of Philosophy at Lincoln University

Abstract

The aim of this study was to refine white clover (Trifolium repens L.) seed crop management. To do this, objectives included to quantify vegetative and reproductive growth and development of white clover grown for seed and quantify the legacy effect of the nitrogen (N) in clover residue following seed harvest and its potential contribution to following crops. Two field experiments were established; one was through conventional cultivation (Conventional) and the other was direct drilled (Direct Drill). Both experiments included the white clover cultivar ‘Romena’, which was grown for seed production. Both experiments were sown on 25 March 2022 at a rate of 3 kg/ha in 30 cm row spacing. There were two treatments: an unsprayed control, and a herbicide canopy management treatment of 1.5 L/ha of Argosy® (25 g/L diflufenican and 250 g/L bromoxynil) and 1.5 L/ha of Relay® Super S (680g/litre 2,4-D ester), applied on 7 September 2022 at approximately growth stage (GS) 33. The chemical canopy management mirrored commercial grower practice and was intended to increase flower bud survival during emergence by opening the canopy, and therefore increase seed yield. The Direct Drill crops received an unplanned application of Gramoxone® (1.5 L/ha, 250 g/L paraquat) on 13 October 2022, which gave a second interruption to canopy photosynthesis. Both experiments were harvested for seed on 30 January 2023 (majority of crop at GS 90+). White clover residue decomposition was measured for 239 days following seed harvest (until 26 September 2023). Sub-plots where plant growth was excluded were used to determine the effect of plant N uptake on soil mineral N. The Conventional clover was replaced with a kale (Brassica oleracea ssp. Acephala) seed crop. The Direct Drill clover was taken for a second year of seed production. The Conventional experiment showed that seed yield was not affected by canopy management (mean 1255 (± 44) kg/ha). Spray treatment plants produced the same number of nodes per stolon (17.6 (± 0.9)) and leaves per plant (68 (± 8)) by the end of peak flowering. This produced no difference in the number of flowers between crops (1,213 (± 74) flowers/m2), with no difference in thousand seed weight (TSW, 0.72 g (± 0.01)). This suggests there were sufficient resources partitioned towards seed fill in all crops. More flowers were lost at immature growth stages (not yet fully pollinated) in the sprayed crops (79% (± 6)) than in the control crops (46% (± 7)) and was attributed to the continued production of leaf canopy in sprayed crops during flowering (0.02 leaves/°Cd (± 0.025)), whereas the control crops had begun senescing leaves (-0.017 leaves/°Cd (± 0.025)). Control crops, which had fulfilled their apparent genetic minimum architecture prior to flowering (78 (± 12) leaves/plant), were able to support their flowers while also partitioning greater dry matter to root biomass. Both Direct Drill treatments demonstrated plasticity in their proliferation following a 300°Cd lag phase from the sprayings, which resulted in more stolons per plant (15.4 (± 1.6)) and subsequently more leaves per plant (281 (± 41)) and flowers produced (34 (± 7) flowers/plant). However, all crops followed a size-dependent reproductive growth pattern and consistently produced 2.1 (± 0.3) flowers/stolon regardless of treatment or establishment method. The severe effects of spraying the Direct Drill crops gave a lower plant population (26 (± 5) plants/m2), which ultimately restricted seed yield (373 (± 45) kg/ha from 556 (± 38) flowers/m2). There was no difference across experiments or treatments in individual flower development. All flowers, regardless of when they emerged during the flowering period, followed the same progression through visual development stages to harvest maturity: 227°Cd (± 16) from bud appearance in the leaf axil until white, a further 84°Cd (± 4) to be half-pollinated and a further 125°Cd (± 9) to be fully pollinated. Flower emergence was bound to the production of its node and subtending leaf. Therefore, any impact on flowering was a result of resource limitation from canopy damage (e.g. 240°Cd/flower (± 42) in Direct Drill sprayed plants during canopy recovery following Gramoxone®), or reduced rate of proliferation due to canopy closure (e.g. a reduction from 44°Cd/flower (± 8.6) prior to canopy closure to 124°Cd/flower (± 60) thereafter). Once peak flowering occurs, the optimum timing of harvest desiccation should be at minimum 440°Cd later to capture the highest population of mature flowers. As canopy management had no effect on total clover dry matter produced (8660 kg DM/ha (± 601) in Conventional, 8590 kg DM/ha (± 370) in Direct Drill), it had no impact on the amount of N available in harvest residue (2.7% N in aboveground biomass). After 14-68 kg N/ha of seed N was removed, both crops had ~230 kg N/ha in clover N above and below ground. Across experiments, the harvest offal contributed an average of 125 kg N/ha and was 70% decomposed within 50 days of seed harvest. This was because the offal was 43% lamina, 24% petiole and 12% florets, which are readily soluble plant components. Once the clover crop was terminated, regrowth, stolons and roots contributed a further ~100 kg N/ha and residue overall took ~1500°Cd to reach 95% depletion. When kale and white clover were growing, total soil N increased by the amount applied as residue N. In the fallow sub-plots, where plant N uptake was excluded, 60-90 kg N/ha was estimated to have been transported in soil solution to below 40 cm depth. Under kale and white clover, the residue N was estimated to have been immobilised into the microbial biomass prior to winter and then mineralised in the spring when soil temperatures increased and the system demanded N. The findings of this research indicated that when herbicides were applied in early September to manage the canopy, there was no increase in flower population from enhanced flower bud survival. Therefore, the practice was of no yield benefit and would only have been worth the cost in applying the herbicide if weed management was needed. The herbicide did not change the crop’s physiology, and only served to disrupt the sink-source balance to the detriment of flowers and canopy capacity to provide for them. The White Clover model in APSIM NextGen was tested and parametrised with Conventional control crop data. This identified the need for more crop phenology to be involved in the prediction of light interception and subsequent biomass production. This research provided the vegetative and reproductive growth and development relationships, according to thermal time, needed to improve crop modelling outcomes. These can provide a base for producing a white clover seed crop model in APSIM that has the capacity to predict aspects of crop phenology. This will improve estimation of biomass production and make progress towards modelling seed production.