Isotope studies of accumulation and cycling of phosphorus and nitrogen below-ground in canola and lupin

Abstract

It is commonly acknowledged that the cycling of nutrients, including phosphorus (P) and nitrogen (N), from plant residues in crop rotations is important for the sustainability of agricultural systems. This is especially the case for Australian low input rain-fed cropping systems, where, due to economic, climatic and edaphic factors, additions of P and N as fertilizers or manures are limited. Optimal management of P and N cycled from break crop residues requires a sound understanding of the quantity of each nutrient in residues and what proportion potentially becomes available for a following cereal crop. A review of the literature (Thesis Chapter 1) highlighted that whilst there is information concerning quantities of N, and to a lesser extent P, contained in mature above-ground crop residues, much less has been reported concerning quantities of P or N of below-ground (BG) residues from various crop species. This is partly because root studies are time consuming and hence expensive to undertake, but also quantification is hampered by the certainty that not all roots can be recovered from soils, especially in fine textured soils. As a result root turnover and nutrient release have largely been investigated under somewhat ‘artificial’ or ‘unrealistic’ conditions - using roots that have been extracted from soil, dried, often chopped or finely ground and finally incorporated back into soil to decompose. More recent innovative studies, summarised in the review (Chapter 1), have used a stem wick-feeding technique to label crop root systems in situ with the 15N isotope. These studies demonstrated that total BG N accumulation for these crop species was larger than quantified from recovered roots alone. The labelling technique allowed for direct in situ quantitative tracing of the N from legume and oilseed root residues into subsequent wheat plants. It was demonstrated that up to 20% of wheat N uptake may be derived from the BG N input by root systems of a previous break crop. The review (Chapter 1) further highlighted that quantitative assessment of the amounts of P accumulated by crop root systems were extremely scarce and there did not appear to be any in situ isotope studies related to P accumulation BG. Hence the work described in this thesis broadly explored the potential to adapt the approaches used for ¹⁵N isotope studies in order to quantitatively assess in situ P accumulation BG by break crop species in soils differing in texture, and the uptake of P derived from those BG break crop residues by a following wheat plant. The specific aims of the work were: i) to adapt the stem wick-feeding technique for use with ³³P to allow in situ quantification of total BG P accumulation by plants, ii) to quantify and compare BG P in two break crops species (an oilseed and a legume) important in Australian rain-fed cropping systems, iii) to assess and measure whether soil texture influences BG P accumulation in canola (oilseed) and lupin (legume), and iv) to trace the fate of break crop BG P relative to BG N in a following cereal (wheat). Preliminary assessment of methodologies used in estimation of BG N in crop plants and their suitability for ³³P studies for BG P were undertaken (Thesis Chapter 2). It was found that the ‘dry’ method frequently used to recover roots for isotope studies (viz: freeze dry manually picked roots with adhering soil, brush roots clean) was comparable to the conventional ‘wet’ root recovery method (viz: washing soil from roots over a sieve), in that similar amounts of root were recovered, which did not differ in P concentration and were not contaminated by soil. Recovery and measurement of roots from field soil cores suggested the amount of P in canola roots in the topsoil (to 0.1m) could be as much as 4 kg ha⁻¹ compared to 1.5 kg ha⁻¹ for rye and less than 1 kg ha⁻¹ for lupin. Other preliminary studies identified that in stem wick-fed plants, ³³P isotope activity was lower where soil P availability (manipulated by P fertiliser addition) was greater. However, the feeding technique could be used to effectively label root systems of lupin with ³³P even at a late vegetative stage of plant growth when it might be considered that the shoot would be the primary sink for P redistributed within the plant. A further study (Chapter 3; Paper 1) confirmed that a substantial proportion (26-51%) of wick-fed ³³P was allocated to recoverable roots of canola and lupin grown in sand. Since this first main study did not detect any ³³P in soil, a mass balance approach was used to determine the amount of unrecovered ³³P, which was suggested to be largely present in unrecovered fine roots, designated as root-derived (RD) P. Using this indirect approach it was estimated that RD P represented 15% of total BG P for canola and 32% for lupin. A subsequent study in deeper pots (Chapter 4; Paper 2) fed a larger amount of ³³P and extended scintillation counting time for samples to improve the method detection limit. This facilitated the direct estimation of unrecovered RD P for canola and lupin at late vegetative stage in two contrasting soil textures, sand and loam. Estimated total BG P accumulation by both crop species was at least twice that of recovered root P and was a greater proportion of total plant P for lupin than canola. There was more unrecovered RD P in the loam than the sand within each species. No ³³P was detected in labile P pools (resin-P or hexanol released-microbial P) at this late vegetative stage of sampling which suggested that there had been no active efflux of ³³P-labelled orthophosphate from labelled roots or any root turnover. However, from a subsequent study (Chapter 5; Paper 3) where ³³P labelled canola plants were sampled at maturity it was evident that after the late vegetative stage root turnover may occur, with 3-5% of fed ³³P detected in the hexanol-released pool and 6-10% in the resin P pool– the higher values being for a loam textured soil which contained a higher proportioned of the fed ³³P than the sand. There appeared to be no translocation of P from roots to shoot between late vegetative stage and maturity since the proportion of fed ³³P recovered BG was the same (70%) at both times. The proportion and amount of canola BG ³³P that was recovered in subsequently grown wheat was higher in the loam (26%; 2.6 mg P) than sand (22%; 1.5 mg P) reflecting the larger pool of BG P in the loam and the faster turnover rate of BG residues. However, this P derived from the previous crop BG residues represented an equal proportion (20%) of the total wheat P uptake in both soils (Chapter 5, Paper 3) since wheat dry matter production was less in the sand. Hence the P benefit from the previous plant BG residues was the same for wheat on both soils. Dual feeding with ³³P and ¹⁵N was used in the final study reported in this thesis (Chapter 6; Paper 4) to simultaneously assess in situ (i) BG N and BG P accumulation by mature lupin and canola, and (ii) the relative contribution from the decomposition of these BG residues to the N and P nutrition of following wheat. The hypothesis tested was that P release from canola BG residues would be relatively greater than from lupin BG residues whereas N release would be relatively smaller. Partitioning of fed ¹⁵N differed from ³³P with the majority of fed ¹⁵N recovered in shoots while a larger proportion of fed ³³P was allocated BG. The amount of total BG P was greater for canola than lupin although lupin had a higher amount of total BG N (75 mg N plant⁻¹) than canola 68 mg N plant⁻¹). C:P ratio of lupin roots was 708:1 and 188:1 for canola. Root C:N ratio was 39:1 for canola and 24:1 for lupin. The N:P ratio for lupin roots was wider (29:1) than canola (5:1), but the N:P ratio of the RD fractions was similar (6:1 canola; 7:1 lupin). Proportion of BG P taken up by wheat was significantly, but only slightly greater after canola (21%) than after lupin (19%), and since BG P was greater for canola this represented 20% of total wheat P uptake and 12% for wheat after lupin. Despite larger lupin BG N, a lower proportion (~8%) was taken up by wheat than from canola BG N (~12%) and so contribution to wheat total N uptake by lupin BG residues (~10%) was surprisingly less than from canola (12.5%). It was concluded from this final study that P uptake by wheat from residues was related to total BG P of the residues but not total BG N. The proportion of P and N from BG residues of mature canola and lupin taken up by wheat did not appear driven by C:P or C:N ratio of recovered roots, but by P concentration of roots, and possibly N:P ratio of BG residues. Research presented in this thesis demonstrates significantly greater amounts of P in BG residues compared to those previously estimated using root recovery methods alone, and that about one-third of total plant P may be partitioned BG. Thus potential P and N benefits to wheat from cycling of break crop root residues are likely to be more substantial than currently thought, and potentially comparable to contributions from an annual P fertilizer addition in low input rain-fed systems. Results further suggest an interaction between release of N and P from BG residues, with an apparent P limitation to the release of N by lupin BG residues; hence C to nutrient ratio of roots was not a good predictor of nutrient release. Lastly, this research also highlights the contribution by root residues of break crops to the longer term fertility of soils, since a large proportion of the BG P and N remains in soil after wheat. In summary, this work develops greater quantitative understanding of the direct contribution of the BG P and BG N of canola and lupin to wheat in terms of P and N supply, and a greater understanding of P and N accumulation in break crop roots. The adaptation of the stem wick-feeding technique for in situ ³³P-labelling of plants opens up exciting future research opportunities in determining the accumulation, fate and interactions of break crop BG P and BG N under undisturbed conditions in following cereals.