Sources of N₂O–N following simulated animal treading of ungrazed pastures

Abstract

It has been previously hypothesised that the treading of pastures by grazing animals can increase nitrous oxide (N₂O) emissions as the result of reduced plant N uptake. In addition, grazing animals urinate and defecate on to soils which can also increase the N₂O emissions. To avoid these additional N inputs, a treading machine was used in two field experiments where the pre-grazing dry matter (DM) was present or removed to test the above hypothesis. The N₂O emissions were measured for 47 and 30 days afterwards, respectively. The soil nitrate (NO₃⁻–N) pool was ¹⁵N labelled prior to treading in Experiment 2. Treading induced greater N₂O emissions and reduced herbage DM yields (by 31%–41%) and soil NO₃⁻–N concentrations when soil gravimetric water contents ranged from 0.45 to 0.63 g water/g soil. A comparison of treading in the presence and absence of pre-grazing DM showed that reduced plant N uptake did not induce greater N₂O emissions. Rather, the ¹⁵N labelling of the NO₃⁻–N pool indicated this pool contributed to the N₂O emissions under treading. In addition, ¹⁵N labelling showed other soil–N pools became available as a consequence of either, (1) treading inducing soil perturbation and releasing N from the soil organic matter–N and/or plant root–N pools, or (2) as a result of simulated grazing causing the release of plant root–N. Soil water soluble carbon (C) also increased under treading in the absence of pasture, supporting the theory that organic–N was released from either the soil organic matter (OM) or plant root pools. Further research should investigate the effect of treading on the turnover and contribution of organic–N pools (roots, OM) in pasture soils in order to more fully understand the contributions these make to background N₂O emissions and effects on N and C cycling in pasture soils.

Keywords:

• ¹⁵N
• compaction
• denitrification
• grazing
• nitrous oxide
• organic matter

Introduction

Anthropogenic emissions of nitrous oxide (N₂O) are of environmental concern because N₂O is both a precursor to compounds involved in stratospheric ozone depletion (Ravishankara et al. 2009) and a greenhouse gas, with a global warming potential of 298 over a 100-year timeframe (Forster et al. 2007). Intensively managed grazed pastures emit N₂O due to regular anthropogenic nitrogen (N) inputs (fertiliser and biological N fixation), animal excreta and animal-induced soil compaction (Oenema et al. 1997). Emissions of N₂O are exacerbated under wet and saturated soil conditions due to reduced aeration (Luo et al. 2008; Balaine et al. 2013). Animal treading can compact soil creating anaerobic soil conditions (Van Groenigen et al. 2005). In the presence of animal excreta, which provides abundant carbon (C) and N substrates, animal treading can enhance N₂O production due to soil compaction (Van Groenigen et al. 2005). Few studies have examined the effects of animal treading on N₂O emissions in the absence of N application to soils via animal excreta or fertilisers. In one of these, Menneer et al. (2005a) showed that N₂O emissions were three to six times higher 8 d after severe treading compared with no treading. Menneer et al. (2005a) reasoned that the higher emissions were due to: (1) compaction of the soil resulting in reduced soil aeration and enhanced denitrification rates; and (2) treading reducing plant growth and thereby reducing plant N uptake, which provided soil microorganisms with more substrate for N₂O production. However, consideration was not given to the possibility that N embodied in the plant tissues was released into the soil as a consequence of leaf burial or root damage during treading, and that such pasture-derived N might have contributed to N₂O emissions. Pasture leaf tissues may contribute to N₂O emissions. For example, plant material harvested, but not ingested, by dairy cattle falls to the soil surface where up to 1% of the embodied N may be released as N₂O (Pal et al. 2012, 2013). The following study was designed to examine the potential for pasture–N to contribute to N₂O emissions following animal treading. It was hypothesised that N embodied in the trodden pasture would contribute to soil inorganic–N pools and the N₂O emissions.

Materials and methods

Experimental site and preparation

Two experimental sites were located in two paddocks at the Lincoln University Dairy Farm (43°38.59′S, 172°26.21′E; elevation 13 m). The soil at both sites was a strongly gleyed, Temuka clay loam (Typic Orthic Gley; [Hewitt 1998]) with impeded subsoil drainage. The pasture species included perennial ryegrass (Lolium perenne L.) and white clover (Trifolium repens L.) which were regularly grazed by dairy cows as previously described (Pal et al. 2012). To avoid the antecedent excreta effects of grazing animals, pasture areas (15 × 20 m) were fenced and animals excluded for six months prior to commencing the experiments. Experiment 1 was performed in paddock 1 beginning in winter (July–October 2010) when soils are wettest and treading effects most likely to be the greatest. This paddock had been sown 3.5 years before the experiment with a mixture of ‘Arrow’ and ‘Alto’ perennial ryegrass cultivars. Experiment 2 was performed in paddock 2 in late autumn (April–May 2011), the paddock having been sown seven years earlier with ‘Bealey’ perennial ryegrass. Otherwise, the two paddocks were identically managed.

Experimental design and treatments

A randomised block design was used in Experiment 1. The four treatments included two levels of herbage (pre-grazing dry matter [DM] present/pre-grazing DM removed subsequently denoted by H1/H0, respectively) and two levels of treading (present/absent denoted by T1/T0, respectively). Each treatment was replicated five times (Fig. S1). Each treatment was imposed on a c. 0.3 m² area equally split between a soil sampling area for measurement of inorganic–N concentrations and pH and an N₂O emissions measurement area (Fig. S1).

For the DM removal (H0) treatment, the herbage was harvested using electric clippers (Oster Shearmaster®, USA) to a height of 1 cm immediately prior to imposing the treading treatments, removed and dried at 65 °C for 48 h to determine kg DM/ha. Removing the DM leaf material to a 1 cm height was performed to reduce the potential for pasture N uptake. Representative subsamples were ground (<200 µm) in preparation for total C and N analyses using standard procedures described by Rowland & Roberts (1994).

For the treading treatment (T1), a mechanical hoof (hoof print area 90 cm²; [Di et al. 2001]) was used. This mimicked the hoof of a Friesian cow which weighed 450 kg, and following Di et al. (2001), applied 220 kPa of pressure on the soil surface at a rate of 377 hoofs/m². The treatment combination H1T0 was undisturbed pasture and hence, the control treatment.

The randomised block design in Experiment 2 had the same treatments as Experiment 1, except the chamber base areas received fertiliser–¹⁵N prior to treading (each replicated five times) and there was an additional ‘control’ chamber where no fertiliser or treading were imposed, giving a total of five treatments (Fig. S2). Fertiliser–¹⁵N was applied in Experiment 2 to distinguish the contributions of soil inorganic–N and plant-derived inorganic–N to N₂O emissions. Thus ¹⁵N-enriched ¹⁵NH₄¹⁵NO₃ (10 atom% ¹⁵N) was applied at the rate of 50 kg N/ha in a water solution (300 mL/chamber) immediately prior to imposing the treatments. The control treatment received an equal volume of deionised water.

One month prior to the start of the experiments, pasture was mown to a height of 10 cm and the cut herbage distributed evenly across the soil surface. Before imposing the treatments, on average, paddock 1 had 1584 ± 110 kg DM/ha (n = 10) on offer. At paddock 2, the corresponding value was not significantly different, averaging 1636 ± 335 kg DM/ha (n = 10). This mass of DM simulated the pre-grazing DM on offer at the Lincoln University Dairy Farm during 2010–11 (Pal et al. 2012).

Soil, herbage and gas sampling and meteorological measurements

Thirty soil cores (7.5 cm depth × 2.5 cm diameter) were collected from each experimental site, prior to the start of each experiment. These were bulked and submitted for analysis to a commercial laboratory (Hill Laboratories, Hamilton, New Zealand). In both Experiments 1 and 2, further soil cores were taken after treatment applications and analysed for inorganic–N and water soluble C (WSC). This occurred on 11 occasions in both Experiment 1 (days 1, 2, 3, 4, 6, 11, 14, 17, 28, 38 and 47) and Experiment 2 (days 1, 2, 3, 4, 6, 9, 12, 13, 16, 23 and 30).

To examine if treading affected plant–N content and to determine the regrowth of trodden pasture, herbage from all plots in both Experiments 1 and 2 was clipped to a 1 cm height, 90 d after treatment application and analysed for total N and C after drying and grinding (Rowland & Roberts 1994). Prior to determining soil inorganic–N, WSC concentrations and inorganic–N ¹⁵N enrichment, a subsample of soil was dried at 105 °C for 24 h to determine the gravimetric soil water content (θ_g). Soil inorganic–N concentrations were determined by shaking the subsampled soil with 2 M KCl in a 1:10 ratio (soil: KCl) on an end-over-end shaker for 1 h followed by centrifugation of the extract at 2000 rev/min (480 g) for 10 min and then filtering (Whatman No. 42). Analyses for ammonium (NH₄⁺–N) and nitrate (NO₃⁻–N) were performed on an Alpkem FS3000 twin channel flow injection analyser (FIA) (Alpkem, College Station, TX, USA) according to Blakemore et al. (1987) using a colorimetric method. For WSC determinations, a subsample of soil was extracted with deionised water in a 1:10 ratio (soil: water) for 30 min on an end-over-end shaker, followed by centrifugation (750 g) of the extractant for 20 min, and filtration (0.45 µm cellulose nitrate membrane filter; [Ghani et al. 2003]) prior to analysis for total organic–C on a Shimadzu total organic carbon analyser TOC 5000A (Shimadzu Oceania, Sydney, Australia) fitted with a Shimadzu ASI–5000A autosampler. Surface soil pH was determined on all gas sampling occasions using a Hanna HI 9025C portable pH meter fitted with a soil surface probe (Broadley-James Corporation, Irvine, CA, USA), after moistening the soil surface with a drop of deionised water. In Experiment 2, ¹⁵N enrichment of the soil inorganic–N pools was determined using the diffusion technique of Brooks et al. (1989) for days 1 to 4 after treatment application. Concentrations of soil inorganic–N after day 4 were below detectable levels for analysis of inorganic–N ¹⁵N enrichment.

Soil bulk densities (five replicates) for all treatments were determined at the end of the experimental period. Soil cores (9 cm diameter × 12 cm deep) were taken from inside the chamber area, θ_g was determined and bulk density calculated, assuming a soil particle density of 2.65 Mg/m³. Soil bulk densities were calculated for depths of 0–3, 3–6, 6–12 and 0–12 cm. Soil temperature (at 10 cm depth), air temperature and rainfall data were obtained from a meteorological station, 1 km away from the field sites.

For soil N₂O emission measurements, headspace chamber bases (45 cm diameter, stainless steel) protruded 12 cm into the soil. These bases contained an annular water-filled trough which provided a gas-tight seal during gas sampling. During gas sampling events, stainless steel chambers (insulated with polystyrene foam; 45 cm diameter, 12 cm high) created a 19.1 L headspace when placed on the bases. For Experiments 1 and 2, soil N₂O flux sampling was performed on 17 occasions over 47 and 30 d, respectively, after treatment application. Samples were taken between 1000 and 1200 h to minimise any effect of diurnal variation (Das et al. 2013; Van der Weerden et al. 2013a).

Gas samples (10 mL) were manually drawn, using glass syringes fitted with three-way taps, and compressed into 6 mL Exetainer® vials (Labco, High Wycombe, UK) at 0, 30 and 60 min, after positioning the headspace cover. Immediately before analysis for N₂O, these gas samples were brought to ambient pressure using a double-ended needle and analysed on a gas chromatograph (8610, SRI Instruments, Torrance, CA) linked to an autosampler (Gilson 222XL, Middleton, WI), as described by Clough et al. (2010). Gas flux calculations were performed using the method of Hutchinson & Mosier (1981). For Experiment 2, 15 mL headspace gas samples were also taken after 3 h. These gas samples were placed into 12 mL Exetainer® vials and equilibrated to atmospheric pressure, immediately prior to determining the N₂O–¹⁵N enrichment, using an automated continuous flow isotope ratio mass spectrometer (CF–IRMS), Sercon, UK (Stevens et al. 1993).

Statistical analysis

Gas flux data were tested for normality using the Anderson-Darling test and if the data were skewed, it was log_e-transformed (ln[flux+1]) to attain normality (Press et al. 1989). The statistical software Minitab version 16.0 (Minitab 2006) was used to perform the analysis of variance (ANOVA) on the flux data to determine if treatment means differed significantly with 95% confidence (P < 0.05). Treatment differences were tested using Tukey's test. All data presented here are mean ± SD (standard deviation).

Results

Experiment 1

Soil properties and meteorological data

Soil chemical properties were suitable for high pasture DM production (Table 1). During the experiment, the highest rainfall (23 mm) occurred on day 38, followed by days 8 and 21 with 4 and 7 mm, respectively (Fig. S3). The average daily air and soil (10 cm depth) temperatures ranged from 1.2 to 14.7 °C and from 2.4 to 9.3 °C, respectively (Fig. S3). Soil θ_g ranged from 0.45 to 0.63 g water/g soil (Fig. S3).

Table 1 Chemical properties of the soil used during the study at paddocks 1 and 2.

Soil properties	Paddock 1 (July 2010)	Paddock 2 (April 2011)
pH (1:2)	6.0	5.9
Total C (g/kg)	39.0	56.0
Total N (g/kg)	4.1	5.5
Anaerobically mineralisable N (µg/g)	162	256
Available N (kg/ha)	198	287
Olsen P (mg/kg)	25	62
Potassium (cmolc/kg)	0.58	0.94
Calcium (cmolc/kg)	10.9	16.1
Magnesium (cmolc/kg)	2.81	3.67
Sodium (cmolc/kg)	0.29	0.51
Cation exchange capacity (cmolc/kg)	20	29
Total base saturation (%)	73	73

Dry matter production, C and N contents

Nitrogen and C contents of the herbage prior to treatment application for Experiment 1 were 37 ± 5 mg N/g DM and 388 ± 9 mg C/g DM (n = 10), respectively. Ninety days after treading, the DM yield was lower (P < 0.05; n = 5) than in the non-trodden plots, averaging 1333 ± 273 and 2257 ± 135 kg DM/ha for the H1T1 and H1T0 treatments, respectively. The DM yields for the corresponding nil-herbage treatments averaged 808 ± 144 and 937 ± 202 kg DM/ha for the H0T1 and H0T0 treatments, respectively, with no treading effect occurring.

Soil pH, soil bulk density and soil gravimetric water content

The treatments had no significant effect on soil bulk density which averaged 0.74 ± 0.14, 1.0 ± 0.10, 1.07 ± 0.05 and 0.97 ± 0.07 Mg/m³ (n = 12) in the 0–3, 3–6, 6–12 and 0–12 cm depths, respectively. Daily average soil pH values ranged from 6.4 to 7.2 (n = 20 on each occasion) and were unaffected by treatment. Under treading, soil θ_g was higher (P < 0.05) with herbage present in the H1T1 treatment (0.56 g water/g dry soil) than in the H0T1 treatment (0.49 g water/g dry soil), while it was 0.52 and 0.66 g water/g soil in the H1T0 and H0T0 (non-trodden) treatments, respectively. Using the mean bulk density of 0.87 Mg/m³, these values translated to soil volumetric water contents of 0.49, 0.43, 0.45 and 0.57 m³ water/m³ soil in the H1T1, H0T1, H1T0 and H0T0 treatments, respectively.

Soil inorganic–N and WSC

On day 1, NH₄⁺–N concentrations were higher (P < 0.05) in the presence of herbage (ranging from 13.5 to 14.8 µg/g dry soil) than in the nil herbage treatments (ranging from 9.8 to 11.4 µg/g dry soil) with no effect of treading. Soil NH₄⁺–N concentrations remained higher (P < 0.05) in the control (H1T0) than the other treatments until day 6 but from day 11 onwards, concentrations did not differ with treatment (Fig. 1A).

Figure 1 Mean soil inorganic–N concentrations under treading and herbage treatments for Experiment 1 (n = 5, error bars are ± SD). Treatments are abbreviated as follows, ‘H1T0’: herbage not clipped and no treading applied; ‘H1T1’: herbage not clipped and treading applied; ‘H0T0’: herbage clipped and no treading applied; ‘H0T1’: herbage clipped and treading applied. A, Ammonium (NH₄⁺–N) concentrations; B, Nitrate concentrations (NO₃⁻–N).

Soil NO₃⁻–N concentrations were higher in the H1T0 treatment during most soil sampling events (Fig. 1B). Treading corresponded with decreased (P < 0.001) soil NO₃⁻–N concentrations throughout the experiment irrespective of the presence or absence of pre-grazing DM (Fig. 1B). Soil NO₃⁻–N concentrations in the H0T1 and H1T1 treatments averaged 3.5 and 3.5 µg/g soil, respectively, while in the H0T0 and H1T0 treatments, the respective values were 9.2 and 10.0 µg/g soil (Fig. 1B). A herbage × treading interaction occurred on days 38 and 47 with higher (P < 0.05) NO₃ -N concentrations recorded in the presence of pre-grazing DM without treading (H1T0).

Over the experimental period, mean soil WSC concentrations were 268 ± 32, 247 ± 33, 226 ± 21 and 223 ± 24 µg/g, in the H0T1, H1T0, H1T1 and H0T0 treatments, respectively (Fig. 2). The WSC concentrations were higher (P < 0.001) on days 2, 4, 6 and 14 (Fig. 2) in the H0T1. Mean WSC concentrations declined after day 1 until day 6, except in the H0T1 treatment, whereupon an increase occurred until day 11, followed by a decline in all treatments (Fig. 2).

Figure 2 Mean soil WSC concentrations under treading and herbage treatments for Experiment 1 (n = 5, error bars are ± SD). See caption of Figure 1 for treatment details.

Nitrous oxide emissions

On day 1, none of the treatments had significantly affected the N₂O emissions (Fig. 3A). However, from days 2 to 14, both treading treatments induced higher N₂O emissions (P < 0.05), more so for the H1T1 treatment than the H1T0 treatment (Fig. 3A). The highest N₂O emissions (P < 0.05) were observed on day 2 and averaged (± SD) 98 ± 68, = 46 ± 37, > 5 ± 3, = 5 ± 12 µg N₂O–N/m²/h for the H0T1, H1T1, H0T0 and H1T0 treatments, respectively (Fig. 3A). Emissions of N₂O under treading treatments were significant (10- to 20-fold higher than control levels; P < 0.05) until day 14 with maximum values on day 2 ranging from 2 to 40 g N₂O–N/ha/d. Cumulative N₂O emissions in the treading treatments were higher (P < 0.05) and ranged from 9.3 to 16.7 mg N₂O–N/m². In the treading treatments, the bulk of the N₂O emissions occurred within the first 14 days, and from days 17 to 47, treatments had no effect on N₂O emissions.

Figure 3 Mean soil N₂O emissions under treading and herbage treatments for Experiment 1 (n = 5, error bars are ± SD). See caption of Figure 1 for treatment details.

Experiment 2

Soil properties and meteorological data

The soil's chemical properties were suitable for high pasture DM production (Table 1). The highest rainfall event occurred on day 30 (16 mm), followed by day 9 and day 10 with 7 and 15 mm, respectively (Fig. S4). The average daily air and soil temperatures ranged from 5.3 to 17.6 °C and 7.9 to 14.5 °C, respectively (Fig. S4). Soil θ_g ranged from 0.50 to 0.59 g water/g soil and fluctuated with rainfall (Fig. S4).

Dry matter production, C and N contents

In Experiment 2, prior to applying treatments, there was 1636 ± 335 kg DM/ha with N and C contents of 33 ± 7 and 402 ± 6 mg/g, respectively, and a ¹⁵N enrichment of 0.367 ± 0.02 atom% (n = 10). Ninety days after imposing treading, DM yields were lower due to treading (P < 0.05; n = 5) averaging 1807 ± 379 and 2608 ± 341 kg DM/ha in the H1T1 and H1T0 treatments, respectively. In the H0T1 and H0T0 treatments, the corresponding values were 990 ± 218 and 1080 ± 261 kg DM/ha, respectively, with no effect of treading. Dry matter ¹⁵N enrichment in the ¹⁵N fertilised plots averaged 1.57 ± 0.25 atom% with no treatment effect, while the ¹⁵N enrichment of the control remained at 0.367 ± 0.01 atom%.

Soil pH, soil bulk density and soil gravimetric water content

The treatments had no significant effect on soil pH which ranged from 6.1 to 7.9. Soil bulk density also did not differ with treatment, at any depth, and averaged 0.69 ± 0.23, 0.99 ± 0.11, 1.05 ± 0.07 and 0.91 ± 0.14 Mg/m³ (n = 12) in the 0–3, 3–6, 6–12 and 0–12 cm depths, respectively. In the H1T0 treatment, θ_g averaged 0.53 g water/g soil (θ_v = 0.45 m³/m³) and this was higher (P < 0.05) than in the H1T1 treatment (θ_g = 0.44 g water/g soil; θ_v = 0.37 m³/m³) on days 2, 9, 12 and 30 due to rainfall. On the remaining days, values did not differ with treatment ranging from 0.54 to 0.60 g water/g soil (θ_v ranged from 0.45 to 0.50 m³/m³).

Soil inorganic–N and WSC

Soil NH₄⁺–N concentrations were higher (P < 0.05) in the H1T0 treatment than in the H0T0 treatment on most days except days 3 and 4 (Fig. 4A). Analysis of ¹⁵N enrichment for NH₄⁺ could not be performed as the soil extract concentrations were too low. Soil NO₃⁻–N concentrations were lower in the H0T1 treatment on most days except days 2, 3 and 9, when concentrations did not differ due to treatments (Fig. 4B). From day 16 onwards, NO₃⁻–N concentrations were lower (P < 0.05) in the H0T1 treatment than the other treatments, and by day 30, concentrations in the H1T1 treatment were higher (P < 0.05) than in the other treatments. The ¹⁵N enrichment of the soil NO₃⁻–N pool was lower (P < 0.001) in the H0T1 treatment on days 1 and 2 (Table 2) when compared with the remaining treatments. On day 4, the lowest ¹⁵NO₃⁻–N enrichment occurred in the H0T1 treatment, with H0T0 and H1T1 having the highest values (Table 2). Treading corresponded with increased WSC concentrations in the absence of DM; the H0T1 treatment had higher (P < 0.001) WSC than in the H0T0 treatment from day 4 until day 12 but treatments did not differ after this time (Fig. 5).

Figure 4 Mean soil inorganic–N concentrations under treading and herbage treatments for Experiment 2 (n = 5, error bars are ± SD). Treatments are abbreviated as follows, ‘Control’: no fertiliser application, herbage not clipped and no treading applied; ‘H1T0’: ¹⁵N fertiliser applied, herbage not clipped and no treading applied; ‘H1T1’: ¹⁵N fertiliser applied, herbage not clipped and treading applied; ‘H0T0’: ¹⁵N fertiliser applied, herbage clipped and no treading applied; ‘H0T1’: ¹⁵N fertiliser applied, herbage clipped and treading applied.

Table 2 Enrichment of ¹⁵N in soil NO₃⁻ in treatments over time.

	Enrichment of NO3^−–^15N (atom%)
Treatment combination	Day 1	Day 2	Day 3	Day 4
Control	0.38 ± 0.03^c	0.38 ± 0.00^c	0.36 ± 0.01^b	0.37 ± 0.03^d
H0T0	2.5 ± 1.5^a	1.9 ± 1.0^ab	2.5 ± 1.6^a	4.0 ± 0.8^a
H0T1	1.0 ± 0.3^b	1.7 ± 0.6^b	2.2 ± 0.2^a	1.2 ± 0.6^c
H1T0	3.3 ± 0.5^a	2.3 ± 0.8^ab	1.5 ± 0.4^a	2.1 ± 0.8^b
H1T1	4.3 ± 0.2^a	3.4 ± 0.2^a	2.1 ± 0.7^a	3.4 ± 0.1^a

Significant differences (P < 0.05) at a given day are shown by different letters in the same column using Tukey's test. Data are mean ± SD. Treatments are abbreviated as follows, ‘Control’: no fertiliser application, herbage not clipped and no treading applied; ‘H1T0’: ¹⁵N fertiliser applied, herbage not clipped and no treading applied; ‘H1T1’: ¹⁵N fertiliser applied, herbage not clipped and treading applied; ‘H0T0’: ¹⁵N fertiliser applied, herbage clipped and no treading applied; ‘H0T1’: ¹⁵N fertiliser applied, herbage clipped and treading applied.

Figure 5 Mean soil WSC concentrations under treading and herbage treatments for Experiment 2 (n = 5, error bars are ± SD). See caption of Figure 4 for treatment details.

Nitrous oxide emissions and ¹⁵N enrichment

Emissions of N₂O from the control were lower (P < 0.05) than the other treatments throughout the experiment (Fig. 6A). Treading corresponded with reduced (P < 0.05) N₂O emissions on days 1, 11 and 12, while the only effect of pasture being present was significantly reduced emissions on days 1 and 12 (Fig. 6A). Cumulative N₂O emissions over 30 d were lower (P < 0.05) in the control treatment while the values from other treatments did not differ significantly from each other averaging 1.9 ± 0.6, 2.0 ± 1.3, 1.6 ± 0.5, 1.6 ± 0.8 and 0.5 ± 0.2 g N₂O–N/m² in the H0T0, H0T1, H1T0, H1T1 and control treatments, respectively.

Figure 6 Mean soil N₂O emissions and ¹⁵N–N₂O atom% under treading and herbage treatments for Experiment 2 (n = 5, error bars are ± SD). See caption of Figure 4 for treatment details. A, Soil N₂O emissions; B, ¹⁵N enrichment of N₂O.

The ¹⁵N enrichment of the N₂O was lowest in the control treatment, ranging from 0.36 to 0.38 atom%, (Fig. 6B). In the H0T0 treatment, the ¹⁵N enrichment of the N₂O generally remained higher (P < 0.05) between day 5 and 30, averaging 4.2 atom% over this period (Fig. 6B). Mean ¹⁵N enrichment of N₂O in the other treatments declined over time from an initial mean of 7.6 atom% (Fig. 6B). Averaged over the experimental period, the ¹⁵N enrichments of the N₂O equated to 5.3, 4.0, 4.3, 4.0 and 0.37 atom% in the H0T0, H0T1, H1T0, H1T1 and control treatments, respectively.

Discussion

As noted above, animal treading may enhance N₂O emissions. In the absence of fertiliser and excreta, Menneer et al. (2005a) observed N₂O emissions that were three- to six-fold higher for 21 d after treading, with a maximum of 52 g N₂O–N/ha/d on day 8. The effect of simulated treading on N₂O emissions in Experiment 1 was comparable.

Reduced DM yields, of 31%–41%, observed under treading in our experiments are in accordance with earlier work showing treading adversely affects shoot and root growth (Cook et al. 1996; Nadian et al. 1996; Drewry et al. 2001; Menneer et al. 2005b; Drewry et al. 2008). Thus, according to the first hypothesis proposed by Menneer et al. (2005a), the reduced pasture growth under treading, observed in Experiments 1 and 2, should have reduced plant N demand and reduced plant competition for soil NH₄⁺–N and NO₃⁻–N. This should have enhanced substrate supply for the microorganisms responsible for generating N₂O emissions. A comparison of the H1T0 and H0T0 treatments allows this hypothesis to be more rigorously assessed, since the latter treatment had pre-grazing DM removed and thus potential ‘plant–N uptake’ was severely reduced. However, in Experiment 1, soil NH₄⁺ concentrations were actually higher when DM was present. Similarly, the soil NO₃⁻–N concentrations in Experiment 1 did not vary between the H1T0 and H0T0 treatments until day 27. In Experiment 2, soil NH₄⁺ concentrations were higher at most sampling times when DM was untrodden. The period of high N₂O fluxes, in Experiment 2, also coincided with a trend for soil NO₃⁻–N concentrations to be higher in the presence of DM (H1T0). This indicates that the removal of plant N uptake, as a consequence of damage to antecedent DM via treading, did not significantly enhance the substrate supply to microorganisms that produce N₂O emissions.

Under treading, DM yields declined, so plant N uptake must have been lower. Assuming that the leaf N concentrations observed in Experiments 1 and 2 on day 0 remained constant after treading, and multiplying by the reductions in DM yield, suggests that treading reduced plant N uptake by 26–34 kg N/ha. Further assuming a soil bulk density of 1.0 Mg/m³ and a soil depth of 0.1 m, the reduced N uptake ranges from 26–34 µg N/g soil or <1 µg N/g soil/d. Thus, if treading increases N₂O emissions, this should not be attributed to reduced N uptake. Based on these arguments, we reject the first hypothesis.

The second hypothesis proposed by Menneer et al. (2005a) attributed higher N₂O emissions following treading to soil compaction inducing a higher denitrification rate. As a corollary, treading should also increase the soil's bulk density. This did not happen in Experiments 1 and 2. However, the changes in θ_g indicate that treading altered the soil's pore size distribution. Most likely, this involved an increased percentage of finer pores which should correspond with reduced soil aeration and increased N₂O production, depending on soil moisture content (Balaine et al. 2013). In Experiment 1, pasture herbage evidently protected the soil from treading damage since θ_g was consistently higher when herbage was present under treading. Conversely, in Experiment 2, θ_g in the H1T1 treatment was lower than in the H1T0 treatment, suggesting the herbage had not protected the soil from treading. The soil moisture contents were sufficient for denitrification to occur (Dalal et al. 2003; Van der Weerden et al. 2013b). Further supporting the second hypothesis of Menneer et al. (2005a) were the lower soil NO₃⁻–N concentrations observed under the treading treatments (Experiments 1 and 2). Based on these arguments, we cannot reject the second hypothesis.

Finally, Experiment 2 also showed that in the presence of herbage, the N₂O–¹⁵N enrichment was not affected by treading. However, when herbage was removed, the N₂O–¹⁵N enrichment was lower with treading than without it. This indicated the ¹⁵N-enriched NO₃⁻–N pool supplying the N₂O production pathways had been diluted by contribution(s) from an antecedent ¹⁴N pool(s). This may have come from soil organic–N or plant root–N, made available as a consequence of treading physically perturbing the soil. Alternatively, plant root–N may have been released in response to cutting of the pasture on day 0 (Macduff & Jackson 1992). Interestingly, WSC levels were highest under treading in the absence of pasture (Experiments 1 and 2) which further suggests treading had promoted either the breaking down of plant roots or soil organic matter. This, and dilution of the NO₃ –¹⁵N pool under treading, also raises a question about how treading influences turnover of organic–N pools (roots, soil organic matter) in pasture soils and the subsequent N transformations.

For dairy cattle grazing, the most susceptible periods when treading damage to soil may occur are autumn, winter and spring, when soil moisture levels are high (Drewry 2006). Thus, poorly drained soils in high rainfall areas are most susceptible, especially if stocking rates are high. Natural recovery of soils from treading-induced soil compaction is a cyclical process that may occur over weeks to years (Drewry 2006). Thus, given typical grazing rotations on dairy farms are in the order of weeks, the most appropriate way to avoid treading inducing N₂O emissions is to prevent it occurring. Luo et al. (2010) suggest the use of a winter stand-off pad could be effective at reducing N₂O emissions from a dairy farm. The results of the current study also show that N₂O emissions under treading are also dependent on inorganic–N concentrations in the soil. Hence, optimising N supply and demand by appropriately timing fertiliser and dairy effluent applications to the soil, to maximise plant uptake prior to treading, would assist in reducing N₂O emissions.

Conclusion

Following simulated animal treading, soil inorganic–N was the main N source for N₂O production. Compaction by animal treading changed the soil's water storage capacity and pore size distribution near the soil surface; the reduced aeration evidently contributed to the increased N₂O emissions. This study supports previous hypotheses that N₂O fluxes following treading are driven by antecedent inorganic–N pools, governed by reduced aeration according to soil wetness. In addition, this study has shown treading induced an N-priming effect with organic matter–N in the soil and/or plant root–N contributing to the N₂O flux.

Supplementary files

Supplementary file 1: Figure 1. Field layout of paddock 1 for Experiment 1 showing the position of the gas chambers and soil sampling plots for the treading and herbage treated plots.

Supplementary file 2: Figure 2. Field layout of paddock 2 for Experiment 2 showing the position of the gas chambers and soil sampling plots for the treading and herbage treated plots.

Supplementary file 3: Figure 3. Meteorological data during the experimental period 1 July–17 August 2010 at paddock 1 for Experiment 1. Rainfall, air temperature and soil temperature (10 cm) are daily average values. Gravimetric water content was measured at 0–7.5 cm soil depth.

Supplementary file 4: Figure 4. Meteorological data during the experimental period 8 April–8 May 2011 at paddock 2 for Experiment 2. Rainfall, air temperature and soil temperature (10 cm) are daily average values. Gravimetric water content was measured at 0–7.5 cm soil depth.

Acknowledgements

Manaaki Whenua Landcare Research, Lincoln, provided funding to the senior author. We are grateful to Shreeparna Pal, Neil Smith, Roger Atkinson and Chris Abraham for assistance in fieldwork and treatment application; and Manjula Premaratne, Qian Liang, Vicky Zhang and Roger Cresswell for GC, inorganic–N, herbage and isotopic analysis, respectively.