The field experiment was conducted over a 12-month period from June 2020 until June 2021 at the University of Queensland, Gatton campus (Gilbert paddock), approximately 85 km west of Brisbane, Australia (27.54 °S, 152.34 °E). The soil is classified as a Black Vertosol using the Australian Soil Classification system (Isbell, 2016) (Table S1). The climate at the site is classified as sub-tropical and is characterized by warm, wet summers and cool, dry winters. Long-term annual average rainfall is 770 mm, received predominantly in the summer months (BoM, 2021). Mean monthly minimum and maximum air temperatures at the site range from 6.3oC (July) to 19.1oC (January), and from 20.8oC (July) to 31.7oC (January), respectively (BoM, 2021).
The experiment was conducted at a site that had hosted crop sequences of maize (2018/19 summer), wheat or faba bean (2019 winter) and maize (2019/20 summer) (Fig. 1). Three replicates of different rates of urea fertilizer had been applied in a randomized complete block design at sowing of the initial maize crop, after which no further N fertilizer was applied during the subsequent winter and summer crop seasons. The initial urea-N rates were 0, 62.5, 125, 187.5 and 250 kg N ha⁻¹, and while there were linear responses to increasing initial N rates recorded in maize biomass production, N uptake, grain yield and grain N removal in 2019/20 (Manandhar et al., 2023), there was no evidence of differential amounts of residual fertilizer N between N rates seen in profile mineral N at sowing of the 2020 winter crops (Fig. S2).
The data reported here are derived from consecutive winter and summer crop growing seasons in 2020 and 2020/21, respectively (Fig. S1), during which wheat-maize or faba bean-maize crop sequences were grown without fresh fertilizer application and sown into plots that had grown the same species the previous winter (i.e., crops 4 and 5 in sequences of M-W-M-W-M or M-F-M-F-M). Following the 2020 winter crop harvest and prior to maize sowing, plots received either standard urea or 15N-enriched urea at rates equivalent to the initial applications in 2018 (i.e., 0–250 kg N ha− 1). Both N fertilizer products were applied at the same rate in separate parts of the same experimental plot, to allow direct comparisons of fertilizer N and legume N recovery by a single maize crop using the NA and EN methods (Fig. 2). The starting mineral N prior to fertilizer N application for the maize crop was 58.4 kg N ha− 1 in faba bean plots and 45.4 kg N ha− 1 in wheat plots (0–60 cm; Fig. S2).
Following the winter crop harvest, maize was sown and fertilized using either standard urea (δ15N, -1‰) or 15N-enriched (5 atom%) urea in treatments applied to spatially separated sections in each treatment/plot. All N fertilizers were applied at five rates (0, 62.5, 125, 187.5 and 250 kg N ha− 1) to determine the shape of the fertilizer N response curve and to track fertilizer N uptake by the 2020/21 summer maize crop. The microplots for 15N-enriched fertilizer were established in clearly marked but spatially separated parts of each treated plot that was otherwise used for NA assessments. This ensured minimum variation in field conditions and easier interpretation of the data while comparing the two isotopic methods (NA and EN) for fertilizer N recovery. The 15N-enriched urea fertilizer was applied manually into a furrow adjacent to a 1.5 m length of one central maize row, with the applied solution immediately covered with the displaced topsoil. The remaining sub-plot area outside the microplot received standard granular (unlabeled) urea fertilizer that was mechanically applied in bands at rates equivalent to that in the 15N-enriched micro-plots. The urea bands (enriched or standard) were applied approximately 0.10 m below the soil surface and offset by approximately 0.20 m from the maize rows (0.75 m row spacing). The layout of the subplot with a 15N treated microplot is shown in Fig. 2.
Irrigations were given to the experimental areas in both seasons, to ensure sufficient soil moisture for crop growth. Weeds and insects were controlled by manual weeding and spraying as required during both growing seasons.
During the initial winter crop season, biomass samples (1 m2) were collected from both the wheat and faba bean crops at maximum biomass, while grain yields were determined from larger, mechanically harvested areas (8.5m * 1.6m = 13.6 m2). During the subsequent maize crop the sampling areas varied depending on the treated areas (e.g., enriched fertilizer 15N microplots versus larger areas receiving commercial urea), necessitating different sampling strategies for biomass and grain production and associated crop N acquisition and partitioning. The timing of biomass sampling was the same in both sub-plots, to allow for easy comparison between methods.
The smaller microplots that received 15N-enriched urea fertilizer (single 1.5 m length row) allowed only a single 1m of crop row (i.e., 0.75 m2) to be sampled close to harvest maturity for determination of biomass and N concentration in grain and stover. Buffer rows either side of the maize row treated with 15N-enriched fertilizer were also sampled to account for foraging of 15N-enriched fertilizer from neighboring crop rows. Grain yields from the treated and buffer rows in the 15N-enriched microplots were determined by manually separating grains from the biomass sampled at physiological maturity.
Biomass samples from the larger plot area receiving commercial urea were collected from twice the sampling area (2 rows x 1m, or 1.5 m2) at a sampling date estimated to represent maximum crop biomass and N uptake before any leaf drop during senescence, while grain yields were determined by harvesting a much larger plot area (10 m x 2 crop rows, or 15 m2) using a small plot harvester (Fig. 2). The mechanical harvesting was undertaken four weeks after biomass sampling once the crop had senesced and grains were dry enough to thresh efficiently.
Powdered biomass, grain and urea fertilizer samples were encapsulated in tin foil cups to form pellets which were analysed for 15N concentration at the UC Davis Stable Isotope Facility (SIF), California, USA, using an elemental analyser interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer (IRMS). The precision of the IRMS was 0.03‰ for 15N.
Nitrogen isotope ratio δ15N (‰) in plant samples was calculated as:
\(\delta 15\text{N} \left(\text{‰}\right)=\left(\frac{\text{R} \text{s}\text{a}\text{m}\text{p}\text{l}\text{e}}{\text{R} \text{s}\text{t}\text{a}\text{n}\text{d}\text{a}\text{r}\text{d}}-1\right)x 1000\) | (Eq. 1) |
where Rsample is the isotope ratio of the sample (15N/14N) and Rstandard is the isotope ratio for the accepted N standard (atmospheric N2, 0.0036765). Multiplying by 1000 expresses the value as per mil (Dalal et al., 2011, Junk and Svec, 1958). Reference materials analyzed with the plant samples at UC Davis lab were glutamic acid (δ15N, -6.8‰), peach leaves (δ15N, 2.0), alfalfa flour (δ15N, 1.8), nylon6 (δ15N, -10.5), bovine liver (δ15N, 7.7), and enriched alanine (δ15N, 41.1).
Representative plant sub-samples were analyzed for total N concentration using the Dumas combustion method on a LECO analyzer at the Chemistry Centre, Department of Environment and Science, Queensland, Australia. Crop N uptake (kg N ha− 1) was calculated by multiplying biomass, stover or grain yield (kg ha− 1) and N concentration (g kg− 1).
The amount of N derived from fertilizer (Ndff) using the NA method (Sanchez and Blackmer, 1988) was calculated as:
\(Ndff\_NA \left(\%\right)=\frac{\left(\delta 15Nf - \delta 15Nuf\right)}{\left(\delta 15Nurea - \delta 15Nuf\right)} x 100\) | (Eq. 2a) |
where δ15Nf = δ15N from the fertilized treatment, δ15Nuf = δ15N from the unfertilized treatment, and δ15Nurea = δ15N of the urea applied in the experiment (i.e., -1.0‰ based on analysis by UC Davis SIF, California, USA).
The amount of N derived from fertilizer (Ndff) using the EN method (Barraclough, 1995) was calculated as:
\(Ndff\_EN\left(\%\right)=\frac{atom\% 15Nexcess plant}{atom \%15Nexcess fertilizer}x 100\) | (Eq. 2b) |
where atom% 15Nexcessplant is atom%15N in the fertilized plant sample minus atom% 15N in the unfertilized plant sample; atom% 15Nexcessfertilizer is atom%15N in fertilizer minus atom% 15N in the unfertilized plant sample; and atom% 15N is the fractional abundance of 15N [15N / (14N+15N)] expressed as a percentage.
\(Ndff \left(kg N ha\_1\right)=\frac{\left[ N uptake \left(kg N ha\_1\right) x Ndff \left(\%\right)\right]}{100}\) | (Eq. 2c) |
where Ndff is estimated using either the NA method (Eq. 2a) or EN method (Eq. 2b); and N uptake refers to biomass, grain or stover N content.
Recovery efficiency of fertilizer N (REN), defined as the percentage of fertilizer N recovered in aboveground crop biomass during the growing season, was derived using the NA method (Chalk, 2018a) or EN method (Hauck and Bremner, 1976, Smith and Chalk, 2018) as:
\(REN \left(\%\right)=\frac{ N uptake (kg N ha\_1)}{Fertilizer N added (kg N ha\_1) }x Ndff\left(\%\right)\) | (Eq. 3) |
where Ndff is estimated using either the NA method (Eq. 2a) or EN method (Eq. 2b), N uptake refers to biomass, grain or stover N uptake.
The proportion of legume N derived from atmospheric N2 by the NA method (%Ndfa_NA) (Chalk and Craswell, 2018) was calculated as:
\(\%Ndfa\_NA=\frac{\delta 15Nreference plant-\delta 15Nlegume}{\delta 15Nreference plant-B}x100\) | (Eq. 4) |
Two variations of this calculation were used. In the first, δ15Nreference plant is δ15N of wheat growing at the site in the same growing season with the same N fertilizer application history, δ15Nlegume is δ15N of faba bean, and B represents the δ15N of the shoots of legumes that are entirely reliant upon N2 fixation for growth. The B value used for faba bean (-0.50‰) was an average value obtained from the literature where faba beans were grown in N-free culture media in numerous glasshouse studies (Peoples et al., 2017, Turpin et al., 2002, Unkovich et al., 2008).
In the second variant, the contribution of the faba bean crop grown in the preceding winter season to the N content of the following summer maize (fixed N in maize by NA method) was determined. In this instance, δ15Nreference plant was the δ15N of maize growing after wheat in the summer growing season, δ15Nlegume was the δ15N of maize grown after faba bean with the same N fertilizer application history. The same B value correction was applied (i.e., -0.50‰) in this instance, given the faba bean legume N source.
The proportion of N in maize (second season) originating from N2 fixed by the faba bean (first season) was also estimated using the EN method (Chalk, 1998, Chalk et al., 1993)
\(Fixed N in maize\_EN \left(\%\right)=(1-\frac{atom\%15N excess of maize after faba bean}{atom\%15N excess of maize after wheat})x 100\) | (Eq. 5) |
where atom%15N excess is atom%15N in the fertilized maize minus atom%15N in the unfertilized maize.
The N derived from soil (Ndfs) in the winter (wheat or faba bean) and summer maize crops was calculated by difference as:
\(Ndfs non legume crops \left(kg N ha\_1\right)=biomass N -biomass Ndff\) | (Eq. 6a) |
where, biomass Ndff = N derived from residual fertilizer applications from the first maize crop.
\(Ndfs faba bean or maize following faba bean \left(kg N ha\_1\right)=biomass N-biomass Ndff-biomass fixed N\) | (Eq. 6b) |
where, biomass Ndff and biomass fixed N were specific to each fertilizer N application rate. The Ndff in the legume treatments was estimated from the Ndff recorded for the non-legume rotation at the same N rate.
Statistical analyses
Data for each experimental season were analyzed using RStudio. Analysis of variance (ANOVA) for a split plot design was used to calculate treatment means, standard errors and significant differences between treatments means (RStudio Team 2018). The structure of the ANOVA was: crop rotation history as the main factor and fertilizer rate as the sub factor. Statistical testing of treatment means separation was done using Fisher-LSD test at p < 0.05. One-way ANOVA and the least significant difference test were used to determine fixed N recovery in the following maize crop and fertilizer N recovery in both rotations.