Introduction

Denitrification is the microbially facilitated process of reducing nitrate (NO3–) and nitrite. (NO2–) to gaseous forms of nitrogen (N), principally nitrous oxide (N2O) as well as environmentally benign dinitrogen (N2). Losses of N2O and N2 to the atmosphere in fertilised cropping systems can represent a substantial loss of applied fertiliser; possibly resulting in reduced plant N uptake and reducing crop N use efficiency (NUE). Furthermore, losses of N2O create environmental concerns; since N2O is not only the third most important long-lived greenhouse gas (after CO2 and CH4), but also largest anthropogenic cause of stratospheric ozone depletion in the foreseeable future (Davidson and Kanter 2014; Ravishankara et al. 2009). Sugarcane is typically produced under conditions that are known to stimulate soil denitrification, i.e. high fertiliser inputs, a wet and warm climate, and N and C inputs due to the increasing adoption of cane trash retention. In contrast to burning, cane trash retention is known to prevent soil erosion (Valim et al. 2016), maintain soil moisture (Ng Cheong and Teeluck 2016; Sandhu et al. 2017), increase soil C (Canellas et al. 2010; Galdos et al. 2009; Thorburn et al. 2012), and provide a source of N (Fortes et al. 2013; Robertson and Thorburn 2007). However, increased soil moisture together with increased C and N substrate availability is conducive for denitrification, and cane retention has been shown to promote N2O emissions (Wang et al. 2016c). Large N2O emissions in sugarcane (Allen et al. 2010; Wang et al. 2016a) are indicative for high denitrification (N2O + N2) rates, and substantial N losses (Thorburn et al. 2017) from sugarcane systems are often attributed to denitrification. For Australian sugarcane cropping systems, only few studies report NUE using labelled 15N fertiliser and suspect denitrification to be responsible for the loss of 25–60% of the applied fertiliser over one season (Chapman et al. 1994; Prasertsak et al. 2002; Takeda et al. 2021; Vallis et al. 1996). However, there is insufficient reliable data on N2 losses from sugarcane soils based on field measurements since it is inherently challenging to measure N2 emissions against the high atmospheric N2 background. The only method that is suitable to directly measure N2 emissions under field conditions is the 15N gas flux (15NGF) method where highly enriched 15N fertiliser is applied to the soil, gas samples are taken using the static chamber method and analysed for their different isotopologues of N2O and N2 via isotope ratio mass spectrometry (IRMS) (Friedl et al. 2020a). One constraint of the 15NGF method is that the 15N label in the soil mineral NO3− pool is subject to dilution, as the nitrification of non-labelled N from the organic N pool leads to a gradual decrease of the 15N label in the soil NO3− pool over time. Consequently, the use of the 15NGF method is restricted to a limited time period following fertiliser (and hence 15N tracer) application, and significant N2 fluxes are often only found 1–4 weeks after tracer application. In sugarcane systems, only two studies exist that measured N2 emissions in the field using the 15NGF gas flux method. Weier (1996) found denitrification losses up to 3.6 kg N ha− 1 day− 1, representing a loss of 40% of the applied fertiliser within 14 days after application from a commercial sugarcane farm in Australia. Warner et al. (2019) measured N2 losses of up to 1.3 kg N ha− 1 day− 1 over a 7-day field campaign using a novel, field-based isotope ratio mass spectrometer system. These large losses highlight that denitrification can be a major loss mechanism from sugarcane soils with possibly significant economic cost to the industry, and strategies to mitigate denitrification losses by fertiliser and crop management are urgently needed.

In order to reduce N losses via denitrification from sugarcane systems, an improved nutrient management is required that better matches crop N demand with supply by adjusting N fertiliser inputs according to the seasonal yield potential, accounting for the ability of soil organic matter to supply N. Another proposed strategy to improve fertiliser NUE and reduce N2O emissions in sugarcane systems is the use of nitrification inhibitors (NIs), delaying the conversion of ammonium (NH4+) to NO3− via nitrification, and thereby reducing substrate availability for denitrifying bacteria. Several studies have shown that the use of NIs can potentially increase crop NUE, reduce N leaching, and mitigate N2O emissions from cropping soils (Abalos et al. 2014; Halvorson et al. 2014; Scheer et al. 2016; Wang et al. 2016b). In a meta-analysis, Akiyama et al. (2010) showed that NIs significantly reduce N2O emissions by 38%, but so far research has focused on N2O mitigation and the combined effect of NI on N2 and N2O emissions remains largely unexplored. Friedl et al. (2017) have shown that the use of the NI DMPP can reduce the N2 losses in subtropical pasture by more than 70%, providing agronomic benefits that can offset the additional cost associated with the use of NIs. However, to date no such data exists for sugarcane cropping systems.

Therefore, the overall aims for this study were to: (i) quantify emissions of N2O and N2 from a subtropical sugarcane system following fertiliser application; (ii) assess the effect of cane trash retention on emissions of N2O and N2; and (iii) evaluate the efficacy of the NI DMPP to reduce N2 and N2O emissions with cane trash retained. As such, this study delivers an improved quantitative process understanding for N2 and N2O emissions from sugarcane soils, testing a NI as a strategy to exploit the benefits of trash retention with minimal environmental offsets.

Materials and methods

Experimental design

The field experiment was conducted on a commercial sugarcane farm at Bundaberg, QLD (24°57′ S, 152°20′ E). The long-term (1959–2012) annual mean temperature in this subtropical region is 21.5 °C (Bundaberg Aero Station, the Bureau of Meteorology, Australia), with the lowest monthly mean temperature in July (16.1 °C) and the highest in January (25.8 °C). Mean annual rainfall is 1027 mm, where over half is received from December to March. The soil is classified as a redoxic Hydrosol (Isbell 2002) or Gleysol (WRB 2015) with loamy sand in the 0–0.3 m layer, underlain by sandy loam at about 0.3–0.6 m depth and sandy clay loam at about 0.6–1 m depth. In the 0–0.3 m layer, total organic C (TOC) and N ranged from 1.00 to 0.80% and from 0.07 to 0.05%, respectively, decreasing to 0.16% TOC and 0.02% N at 0.6−1 m depth. Soil pH ranged from 6.0 in the topsoil (0–0.3 m) to 5.7 at depth (0.6 − 1 m). Detailed information on selected soil physical and chemical properties is given in Table S1.

Sugarcane (CV Q238) was planted in the middle of raised beds (ca. 1.26 m wide) with a row spacing of 1.83 m in September 2013. The plant cane crops were fertilised at 145 kg N/ha, 28 kg P/ha, 100 kg K/ha and 25 kg S/ha by placing the fertilisers in bands (ca. 0.1 m under the surface) on both sides of the cane setts (ca. 0.05 m away). Following harvest of the plant cane in September 2014, three treatments were applied to compare DMPP-urea with conventional urea and assess the effect of the sugarcane trash blanket (cane residue retention):

  1. (i)

    DMPP: Urea fertiliser with 0.6% DMPP solution (w/w) and sugarcane trash retained in the field.

  2. (ii)

    Trash retained (TRT): Standard Grower Practice, Urea fertiliser and sugarcane trash left in the field.

  3. (iii)

    Trash Removed (TRM): Urea fertiliser with sugarcane trash removed from the surface.

The treatments were arranged in a randomised block design with four replicates. The plots were 6.7 m along the row length and 9.2 m across five rows. A steel base (0.22 × 0.22 m) was installed on one side of the middle row in each plot for the manual chamber measurements. 15N urea (60 atom %) was applied in solution corresponding to the recommended N fertiliser rate of 145 kg of N ha− 1 as a band in the middle across the steel base, buried 0.1 m deep and ~ 0.05 m the sugarcane row. In the TRT and DMPP treatments a 0.05 m thick layer of sugarcane trash was placed on the soil surface, consistent with sugarcane trash management in the growing region.

Gas sampling

The 15N gas flux method was used to measure N2 and N2O emissions by quantifying the increase in 15N-labelled gases in the chamber headspace over time as described by Friedl et al. (2017). Gas samples were taken manually from the chambers once per week for 85 days after fertiliser application, except for the first week of the experiment when samples were taken twice. Polyethylene chambers with a headspace height of 0.314 m were placed on the steel frames, ensuring airtight conditions. Headspace gas samples (20 ml) were taken by connecting a syringe to a 2-way luer-lock tap installed in the lid of the chamber. Gas samples were then injected into a pre-evacuated 12 ml glass vial with a double wadded Teflon/silicon septa cap (Labco, UK). Headspace gas samples were collected at 0, 1 and 3 h after closure from each chamber. In addition, the ambient and soil temperature was taken once per replicate and hour and used to correct flux calculations for temperature.

Gas analysis and N2O and N2 flux calculations

All gas samples, taken 0, 1 and 3 h after closure, were analysed for N2O by gas chromatography (GC) (Shimadzu GC-2014). Following GC analysis, gas samples taken at 0 and 3 h were analysed for the isotopologues of N2 (15N14N, 15N15N) and N2O ([14N15N16O + 15N14N16O] and 15N15N16O) using an automated isotope ratio mass spectrometer (Sercon, 20–20, UK) linked to a Sercon Cryoprep trace gas concentration system.

Flux rates of N2O were calculated from the slope of the linear increase in gas concentration during the closure period. The coefficient of determination (R2) was used as a quality check for linearity and flux rates were set to 0 if R2 was < 0.80. Flux rates were corrected for temperature, air pressure and the ratio of cover volume to surface area as described by Scheer et al. (2014).

The ion currents (I) at m/z 44, 45, and 46 enabled the molecular ratios 45R (45I/44I) and 46R (46I/44I) to be calculated for N2O, and I at m/z 28, 29 and 30 enabled 29R (28I/29I) and 30R (28I/30I) to be calculated for N2. Fluxes of N2 were calculated using the equations given by Spott et al. (2006), calculating the enrichment of the NO3− pool undergoing denitrification ap, and the fraction of N2 derived from this pool (fp):

$${f}_{p}=\frac{{a}_{m}-{a}_{bgd}}{{a}_{p}-{a}_{bgd}}$$
(1)

where abgd is the 15N abundance of the atmospheric background and am is the measured 15N abundance of N2 from headspace gas samples taken 0 and 180 min after closure, respectively. Both abgd and \({a}_{m}\) are calculated as

$${{a}_{i }}_{ }=\frac{{}^{29}R+2\text{*}{}^{30}R}{2\text{*}(1+{}^{29}R+{}^{30}R)}$$
(2)

The 15N enrichment of the soil NO3− pool undergoing denitrification ap is calculated for N2 and N2O (ap N2 and ap N2O) as

$${a}_{p}=\frac{{}^{30}{x}_{m}-{a}_{bgd}*{a}_{m}}{{a}_{m}-{a}_{bgd}}$$
(3)

The measured fraction of m/z 30 in N230xm is calculated as:

$${}^{30}{x}_{m}=\frac{{}^{30}R}{(1+{}^{29}R+{}^{30}R)}$$
(4)

To calculate ap N2O, 45R and 46R were converted to 29R and 30R by correcting for the naturally occurring O2 isotopes:

$${}^{29}R={}^{45}R-{}^{17}R$$
(5)
$${}^{30}R={}^{46}R-{}^{29}R*{}^{17}R-{}^{18}R$$
(6)

Using the value of 0.00038 for 17R and 0.002079 for 18R (Arah 1997). If only 29R was > the detection limit (DL), fp was calculated as

$${f}_{p=}\frac{1}{1-\frac{{}^{29}R{(1-{a}_{p})}^{2}-2{a}_{p}(1-{a}_{p})}{{}^{29}R{(1-{a}_{bgd})}^{2}-2{a}_{bgd}(1-{a}_{bgd})}}$$
(7)

using ap derived from N2O assuming N2 and N2O evolving from the same NO3− pool undergoing denitrification. The headspace N2 concentrations, corrected for air pressure and temperature, were multiplied by the respective fp values giving N2 fluxes expressed in g N2–N emitted g− 1 soil day− 1. Cumulative fluxes of N2O and N2 were calculated by linear interpolation between sampling events.

The between batch precision of the IRMS for N2 based on the standard deviation of atmospheric air samples (n = 60) at 95% confidence intervals (Friedl et al. 2020a) was 1.46 × 10− 6 and 1.36 × 10− 6 for 29R and 30R, respectively. The corresponding method detection limit (DL) ranged from 5.5 mg N2–N m− 2 day− 1 with ap assumed at 60 atom % to 16.6 mg N2–N m− 2 day− 1 with ap assumed at 20 atom %. If N2 fluxes were below the DL, fluxes were set to 50% of the calculated DL.

The fraction of fertiliser derived N2O and N2 (ndff) was calculated as the ratio of 15N atom excess % of N2O or N2 and the 15N atom excess % of the N fertiliser applied, with the soil derived fraction calculated as the difference of 1 and ndff. The measured 15N atom fraction in N2O (a’s) was used for the 15N recovery in N2O, accounting for N2O from nitrification and denitrification mediated pathways. For N2, ap N2 was used for the 15N recovery in N2. If only R29 > DL, ap derived from N2O was assumed as the enrichment of the NO3− pool undergoing denitrification and used to calculate 15N in N2.

Auxiliary measurements

Soil water content in the topsoil (0–0.1 m) was recorded using in-situ moisture probes and a data logger (ThetaProbe, Delta-T Devices Ltd, UK). Soil samples (0–0.1 m and 0.1–0.3 m ) for soil mineral N analysis were taken directly from the fertiliser band next to crop row 2 or 4, twice during the experimental period, 9 and 59 days after fertilisation. Three soil samples were extracted for NO3− and NH4+ with 2 M KCl (1:5 w:v) for 1 h on a rotary shaker, filtered through Whatman no. 40 filter papers and analysed colorimetrically (Rayment and Lyons 2011).

Statistical analysis

Statistical analyses were conducted with SPSS 27.0 (SPSS Inc.,2020). Treatment effects on N2, N2O, and soil mineral N concentrations were examined by analysis of variance (ANOVA) (p < 0.05). Normal distribution of the data was assessed by the Shapiro-Wilk test for normality. Homogeneity of variance was verified by Levene’s test for equality of variances. Differences between treatments were assessed using the Ryan-Einot-Gabriel-Welsh test. Effects of treatments on the temporal pattern of a’s were examined using linear mixed effect models accounting for repeated measurements by specifying chambers as subjects with repeated measures over time. Values in the figures represent means ± standard error of the mean.

Results

Emissions of N2 and N2O

Of 168 N2 fluxes, 15% were discarded due to analytical problems, and 78% exceeded the DL, of which 36% were calculated using ap derived from N2 and 64% were calculated using ap derived from N2O.

Significant emissions of N2 could be measured in all treatments and ranged from 33.4 to 73.4 mg m− 2 day− 1. In the week before N fertilisation, 93 mm of rain fell at the site. The temporal pattern of N2 emissions showed elevated N2 emissions in all treatments on day 1 after fertiliser application followed by 42 mm rain within 2 days, and a declining trend over the first 3 weeks of the experiment with little rainfall and declining soil moisture levels (Fig. 1). After a period of heavy rainfall and increasing soil moisture levels, a second peak of N2 emissions was observed on day 44 of the experiment in all treatments. From day 50 onwards, only small fluxes (< 10 mg m− 2 day− 1) were observed in the DMPP and TRM treatments, with slightly higher N2 emissions in the TRT treatment. Over the 85 days observation period cumulative N2 losses from the fertiliser band amounted to 1703.44 ± 212.74 mg N2–N m− 2 in the TRT treatment with significantly lower losses (< 1200 mg N2–N m− 2) in the DMPP and TRM treatments (Table 1).

Fig. 1
figure 1

Average daily N2 and N2O fluxes for for the different treatments a, b and c with the corresponding product ratio N2O/(N2 + N2O) (d), soil water filled pore space (WFPS) 0-10 cm depht, (e) and daily precipitation (f)

Full size image
Table 1 Cumulative emissions of N2 and N2O from the fertiliser band in a subtropical sugarcane system over 85 days and their respective soil and fertiliser derived fractions
Full size table

Emissions of N2O showed a distinctly different temporal pattern than the N2 fluxes (Fig. 1). On day 1 after fertiliser application only very small fluxes were observed in all treatments. From day 4 onwards significantly elevated N2O emissions were observed with highest emissions in the TRT treatment (116.9 mg N2O–N m− 2 day− 1 on day 8). N2O emissions stayed elevated for approximately 4 weeks and returned to very small fluxes (< 0.5 mg N2O–N m− 2 day− 1) from day 36 onwards in all treatments. Over the 85 days, 1913.0 mg N2O–N m− 2 and 936.4 mg N2O–N m− 2 were lost in the TRT and TRM treatments, respectively, with significantly reduced losses of 48.7 mg N2O–N m− 2 in the DMPP treatment.

Changes in the corresponding N2O/(N2 + N2O) product ratio are shown in Fig. 1. Values across all treatments showed a high variation and ranged from 0 to 0.89. In the TRT and TRM treatments the N2O/(N2 + N2O) ratio showed a clear peak from day 4 to day 36 after fertiliser application. During this period the N2O emissions accounted for 67% and 60% of the total N2 and N2O losses, in the TRT and TRM treatments, respectively. Only DMPP treatment showed constantly low N2O/(N2 + N2O) ratios (ranging from 0 to 0.16), due to very small N2O fluxes coinciding with elevated N2 emissions. Over the 85 days experimental period N2O losses accounted for 46–48% of the total N2 and N2O losses in the TRT and TRM treatment, with a significantly lower N2O/(N2 + N2O) ratio of 4% in the DMPP treatment (Table 1).

Total denitrification losses (N2 + N2O) were highest in TRT (36164.1 mg m− 2), followed by TRM (20608.8 mg N m− 2) and DMPP (11548.7 mg N m− 2) (Table 1), with 75–81% of the N2 losses and 23–51% of the N2O losses derived from fertiliser (Table 2). This represents a loss of 7%, 4% and 2% of the fertiliser N applied as a band from the TRT, TRM and DMPP treatment respectively.

Table 2 Soil and fertiliser derived fractions of cumulative N2 and N2O emissions from the fertiliser band in a subtropical sugarcane system over 85 days
Full size table

The temporal variation of the 15N fraction in N2O differed between treatments, showing values around 0.5 for the 15N atom fraction in N2O for TRT and TRM in the beginning of the experiment, decreasing sharply 36 days after fertilisation to less than 0.005 for the rest of the experiment (Fig. 2). The initial 15N fraction of N2O in the DMPP treatment was below 0.3, decreasing to < 0.005 until the end of the experiment. Values for ap N2O and ap N2 ranged from 0.24 to 0.56, and 0.27 to 0.59, respectively and agreed largely over the time of the experiment (Figure S1).

Fig. 2
figure 2

The15N atom fraction in N2O (a’s) for the different treatments

Full size image

Soil mineral N content

Nine days after fertilisation soil NO3− levels ranged from 19.1 to 47.1 NO3− mg kg− 1 soil in the 0-0.1 m layer, and 18.4 to 44.0 NO3− mg kg− 1 soil in the 0.1–0.3 m layers with lowest concentration in the DMPP treatment, although with no significant difference between the treatments due to a high variability between the replicates (Table 3). Soil NH4+ concentration 9 days after fertilisation ranged from 53.2 to 97.2 NH4+ mg kg− 1 soil in the 0-0.1 m layer, with highest average values in the DMPP treatment, but no significant treatment effect. Soil NO3− and NH4+ levels dropped to almost zero in all treatments 59 days after fertiliser application, with slightly elevated levels in the DMPP treatment in the 0-0.1 m layer.

Table 3 Soil mineral N (NH4+ and NO3−) at 0–0.1 m and 0.1–0.3 m depth in the fertiliser bands 9 and 59 days after fertilisation for the different treatments including Urea, DMPP and Trash removed
Full size table

Discussion

Novel fertiliser management strategies are needed to reduce environmental pollution, mitigate climate change, and increase profitability in sugarcane farming systems. Such strategies need to target denitrification, as wet and warm climatic conditions in combination with high fertiliser inputs and increasing crop residue retention are especially conducive for high losses of fertiliser N via this pathway. The first field study to investigate the effect of the NI DMPP and sugarcane trash retention on N2 and N2O losses from a commercial sugarcane farm demonstrates (a) an increase of denitrification by sugarcane trash retention, (b) high losses of N2 and N2O from banded N fertiliser, (c) a substantial reduction of total denitrification losses (N2 + N2O) from the fertiliser band by DMPP and (d) a shift in the N2O:N2 ratio towards N2 by DMPP.

Temporal variability of N2O and N2 emissions across treatments suggests that NO3- concentration around the N fertiliser band determines the N2O/(N2+N2O) ratio

Fertiliser N application followed by 42 mm rain resulted in highly elevated N2 emissions at day one from all treatments. These emissions (33.4–73.4 mg m− 2 day− 1) are within the range of N2 emissions previously reported from sugarcane systems after the application of N fertiliser (Warner et al. 2019; Weier 1996) and show that the fertiliser applied as urea was rapidly converted to NO3− within the soil and available for denitrification to take place. Interestingly, there were only very small N2O emission on day 1, indicating that the NO3− was completely reduced to N2 during the denitrification process. This is in contrast to other studies that investigated N2 and N2O fluxes from fertilised cropping system in the field where often a burst of N2O emission on day one and a shift towards full denitrification to N2 over time is reported (Buchen et al. 2016; Friedl et al. 2017; Warner et al. 2019). This pattern has commonly been explained by increasing anaerobic condition due to increased O2 consumption in the soil profile (Meyer et al. 2010; Rohe et al. 2021; Yang et al. 2021) and/or the time required for the activation of the N2O reductase activity responsible for the reduction of N2O to N2 (Morley and Baggs 2010; Zheng and Doskey 2016). The different pattern in our study can most likely be explained by > 90 mm rainfall in the week prior to the experiment and its effect on denitrification enzyme activity, together with the dynamics of N substrate supply from the urea band. The high N2 emissions at day one clearly show that reduction enzymes involved in the denitrification process including the N2O reductase were already activated, likely due to anoxia in the soil caused by prior rainfall (Uchida et al. 2014), promoting a complete reduction of NO3− and high emissions of N2 in all treatments. However, it seems that increasing concentrations of NO3− around the fertiliser band inhibited either de novo synthesis or activity of the N2O reductase increasing N2O emission and the corresponding N2O/(N2 + N2O) product ratio. The low N2O emission and N2O/(N2 + N2O) product ratios in the DMPP treatment however indicate inhibition of nitrification, preventing a build-up of NO3− and thus promoting the reduction of N2O to N2. This line of argument agrees with the reported suppression of the N2O reductase gene nosZ by high NO3− concentrations, and the attenuation of this effect by DMPP, leading to an increase in nosZ abundance and a reduction in N2O emissions (Friedl et al. 2020b). The decrease of N2O and N2 emissions in all treatments between day 8 and day 29 can be attributed to a gradual decrease of mineral N concentration in the fertiliser band. It seems that in particular the period of heavy rainfall between day 29 and day 50 after fertiliser application resulted in a depletion of soil mineral N via plant uptake, leaching and denitrification. The simultaneous decrease in the 15N enrichment of N2O (a’s) in the TRT and TRM treatments (Fig. 2) is consistent with consumption of NO3- via these pathways, as co-occurring and/or following nitrification leads to a dilution of the 15N label in the soil NO3- pool.

A second peak of N2 emissions was observed in all treatments on day 44 of the experiment. This peak occurred after a series of heavy rainfall events leading to increased soil moisture, saturation and thus promoting denitrification in the soil profile. In contrast to N2, N2O fluxes showed no significant response to this rainfall event, resulting in almost complete denitrification to N2 with N2O/(N2 + N2O) product ratios ranging from 0.001 to 0.003. This can be most likely attributed to the combination of limited O2 availability and increased entrapment of N2O in the soil matrix after rainfall, together with depleted levels of soil NO3− more than 6 weeks after fertilisation. The low N2 and N2O fluxes after day 50, despite the occurrence of several heavy rainfall events show, that denitrification was limited by N substrate availability, which is also reflected in the low mineral N values measured at day 59 in all treatments.

DMPP reduces magnitude and N2O/(N2 + N2O) ratio of N2O and N2 emissions from the fertiliser band

The use of DMPP lowered N2 emission by 35% and N2O emission by 98% compared to the standard TRT treatment, respectively. This is among the highest reported N2O reductions efficiency of an NI. In a recent meta-analysis Ruser and Schulz (2015) summarised the results from 140 field studies on the effect of different NIs on N2O emissions from agricultural soils and found a mean reduction potential of approx. 35%, but a wide range of inhibitory effects has been reported by numerous studies depending on soil type, climate and management system. The high N2O reduction agrees well with an inhibitory effect of DMPP ranging from 60 to 83% in sub-tropical summer cereal cropping system in Australia where the fertiliser was applied as a band besides the crop rows (De Antoni Migliorati et al. 2016; Scheer et al. 2016), indicating that NI might be especially effective in reducing N2O emissions from fertiliser bands. Banding N-fertilisers creates a biochemical environment in which different N transformations will occur, compared to when the same product is broadcast or mixed into soil which has implications for the effective utilisation of NIs when applied in sub-surface fertiliser bands (Janke et al. 2020). A  sub-surface fertiliser N band will create localised zones of high NO3− concentration that are particularly vulnerable to denitrification. Our results indicate that DMPP is very efficient at inhibiting the conversion of urea into NO3− from the fertiliser band leading to the high reduction in N2O emissions. This in contrast to observations by Janke et al. (2019) that did not find enhanced inhibition of nitrification with DMPP in a highly concentrated urea band in an soil incubation essay, highlighting the difficulties to transfer results from small scale soil incubations to the field level. As incubation containers restrict diffusion of N from the simulated band, high concentrations of ammonia following urea hydrolysis can exert inhibitory effects on the second step of nitrification, i.e. the oxidation of NO2– to NO3− (Breuillin-Sessoms et al. 2017). These conditions may be representative for the urea band itself, but not for zone around the band, where diffusion supplies N that is readily nitrified. This assumption is consistent with the immediate onset of fertiliser derived N2O and N2 emissions in the study presented here, demonstrating rapid supply of NO3− shortly after urea application, effectively limited by the use of DMPP.

The effectiveness of DMPP in our study is further supported by the significant reduction of N2 emission in the DMPP treatment and the observed shift of denitrification losses towards N2. It is known that during denitrification, NO3− competes with N2O as terminal electron acceptor, leading to high N2O/(N2 + N2O) ratios under high NO3− concentrations, and DMPP has been shown to mitigate this effect (Friedl et al. 2020b). However, there is hardly any field data available that investigated the effect of DMPP on both N2 and N2O emission under field conditions. Friedl et al. (2017) reported that DMPP reduced N2 losses from a sub-tropical pasture by more than 70%, providing the first field-based evidence that DMPP can substantially reduce N2 emissions, but found no effect on N2O emissions. The contrasting finding that in our study DMPP had the strongest reduction effect on N2O can be explained by the different fertiliser management and soil N cycling in sugarcane systems. In our study, a high rate of N fertiliser was applied in a band, leading to a build-up of NO3− in the soil, while at the pasture sites a low rate of fertiliser was broadcasted at the surface which, combined with tight N cycling between the soil microbial biomass and plants in pasture soils, limited NO3− concentrations resulting in complete denitrification to N2. Overall, our study suggests that the use of DMPP in sugarcane systems with banded fertiliser does not only offer environmental benefits by reducing N2O emissions but also substantially reduces overall denitrification losses, supporting findings from a modelling study highlighting DMPP as an economically viable strategy to improve NUE and reduce N2O emissions for Australian sugarcane industry (De Antoni Migliorati et al. 2021).

Cane trash management has no effect on the N2O/(N2 + N2O) ratio

Cane trash removal reduced N2 emission by 34% and N2O emission by 51%, respectively. Numerous studies have reported a stimulation of N2O emissions in response to crop residue retention and a recent meta-analysis found an average increase in soil N2O emissions by 43% with crop residue incorporation compared to residue removal (Abalos et al. 2022). This average increase agrees well with the results of our study, but it needs to be noted that the effect of crop residue retention on N2O emissions depends on multiple factors including the C:N ratio and the amount of residues returned, mineral N content of the soil, and the increase of soil moisture due to residue retention. Overlapping effects of these factors can lead to both increases and reductions in N2O emissions due to residue retention (Nguyen et al. 2015). In sugarcane systems overall a stimulation of N2O emissions with cane trash retention has been reported (Wang et al. 2016c). The study presented here shows for the first time an increase of both N2O and N2 emissions in response to cane trash retention. Residues may supply N for the production of N2O and N2 despite their high C:N ratio, yet absolute amounts in comparison to those from the urea band are likely to be negligible. The temporal pattern of soil water content suggests slightly higher WFPS in the trash retained treatments (TRT, DMPP), which may explain at least partially increased emissions of N2 and N2O as compared to trash removed. The trash blanket is also likely to release labile C compounds upon wetting, which is consistent with previous research that reported increased rates of soil denitrification in the presence of easily degradable C compounds (Senbayram et al. 2012; Weier et al. 1993), linked to the creation of anoxic micro-sites in the soil profile (through increased oxygen consumption via soil respiration) and the supply of an energy source to denitrifying microorganisms in the soil. However, the effect of available organic C in soils on the N2O/(N2 + N2O) product ratio is still not fully understood, but a shift in the N2O:N2 ratio towards N2 is generally assumed (Giles et al. 2012; Li et al. 2021). In our study, there was no significant effect of trash management on the N2O/(N2 + N2O) product ratio with almost identical ratios with or without trash retention (TRT, TRM) over the entire experimental period. These findings, together with the reduction of the N2O/(N2O + N2) ratio by DMPP indicate that the ratio was mainly controlled by the high NO3− availability around the fertiliser band, confirming results from a previous study that found substantially lower N2O/(N2O + N2) product ratio after the application of labile C only in soil with low NO3− contents (Senbayram et al. 2012).

Sources of N2O and N2

The enrichment of the NO3− pool undergoing denitrification ap was calculated from N2O, and whenever possible, from N2 data. Comparing ap derived from N2O and N2 shows a good agreement between both, supporting the assumption of a relatively uniform 15N labelled NO3− pool (Friedl et al. 2020a). Interestingly, the DMPP treatment had no effect on ap, despite the decrease in N2O emissions. This suggests that DMPP reduced the nitrification of both fertiliser as well as soil N, which likely led to a reduced NO3− pool for denitrification, but had no effect of the proportion of 15N labelled (fertiliser derived) and non-15N labelled (soil derived) N, resulting in similar ap values as compared to TRT and TRM. This assumption is supported by previous studies demonstrating reduced nitrification of soil derived N (Friedl et al. 2017), as rainfall and/or irrigation after N fertilisation promote the diffusion of both N fertiliser and NIs, leading to a decoupling of N fertiliser and NIs. Denitrification is assumed as the sole source of N2, and consequently DMPP showed no effect on the soil and fertiliser derived fraction of N2 emissions (Table 2). Even though ap showed no treatment effect, the overall 15N fraction in N2O a’s was lower in the DMPP treatment (Fig. 2) compared to TRT and TRM. Furthermore, DMPP increased the soil derived fraction of emitted N2O. The NI DMPP targets the ammonia monooxygenase (AMO), the enzyme catalysing the first step of nitrification, the conversion of NH4+ to hydroxylamine (NH2OH). However, not all nitrifiers oxidise NH4+ via AMO (Martikainen 2022; Wood 1990), and N2O production pathways including heterotrophic nitrification by fungi may not be affected by the use of DMPP. The lower a’s values in the DMPP treatment suggest a larger relative contribution of N2O production pathways bypassing AMO when classic autotrophic nitrification is inhibited, consistent with the larger contribution of soil derived N to N2O emissions in the DMPP treatment. The significant reduction of N2O by DMPP suggests that nitrification in this sugarcane soil is dominated by nitrifiers which possess AMO. Nevertheless, research under controlled conditions is needed to evaluate N2O source partitioning in response to NIs from sugarcane systems, as the relative contribution of heterotrophic/fungal N2O production pathways may be of significance for overall N2O emissions in these agro-ecosystems with residue retention.

Conclusion

The first field study investigating the effect of the NI DMPP and sugarcane trash retention on N2O and N2 losses from a sugarcane system demonstrates significant losses of N2O and N2 from a subsurface N fertiliser band, with more than 50% of these gaseous N losses emitted as N2O. The large amount of N2O lost highlights that high N substrate availability close to the fertiliser band is conducive for denitrification losses, and that ensuing high NO3− concentrations shift the N2O:N2 ratio towards N2O. Emissions of N2O and N2 in this study were measured from the N fertiliser band, likely accounting for a large fraction of overall N2O and N2 loss. Nevertheless, further method development is needed to measure/upscale N2O and N2 emissions from both fertilised (band) and non-fertilised (furrow) areas, as current analytical methods are limited to fertilised areas within a field. Trash retention increased the magnitude of N2O and N2 emissions likely reflecting overlapping effects of increased soil water content and labile C supply from residues. The lack of response of the N2O/(N2 + N2O) ratio to trash management however suggests that soil NO3− concentrations predominantly control the N2O/(N2 + N2O) ratio in banded systems with trash retention. The NI DMPP was not only extremely effective in reducing overall N2O and N2 losses but also in promoting complete denitrification of N2O to environmentally benign N2, with only 4% of total N2O and N2 losses emitted as N2O. This shows that DMPP might be especially effective in reducing N2O emissions from banded fertiliser were localised zones of high NO3− concentration around the fertiliser band can be expected. Consequently, the use of DMPP appears to be an effective strategy to minimise N losses, while keeping the benefits of cane trash retention. Assessing DMPPs effect on other N loss pathways and sugarcane yield will show if the reduction in overall N losses allows for lower N fertiliser rates. Combining NIs with reduced N rates compensates for the price premium of DMPP fertilisers, making it an economically viable strategy to improve NUE while reducing the greenhouse gas-footprint of the sugarcane industry.