Nitrate leaching losses from lysimeters simulating winter grazing of fodder beet by dairy cows

ABSTRACT

Nitrate-N (NO₃⁻-N) leaching losses from dairy cow urine and non-urine areas of simulated winter grazing of fodder beet (Beta vulgaris L.) were quantified using large intact soil monolith lysimeters (50 cm diameter × 70 cm depth), containing Balmoral/Lismore stony silt loam soil. Fresh cow urine was applied to half of the lysimeters in mid- to late-June at 300 kg N ha⁻¹ in 2012 and 250 kg N ha⁻¹ in 2013, following simulated fodder beet grazing. The remaining half did not receive urine to represent non-urine areas of a grazed paddock. Most of the N leached was from the urine-treated lysimeters and 92%–98% was in the form of NO₃⁻-N. Total NO₃⁻-N leached from the urine-treated lysimeters represented an equivalent of 21% (64 kg NO₃⁻-N ha⁻¹) and 32% (84 kg NO₃⁻-N ha⁻¹) of the total urine-N applied in 2012 and 2013, respectively. Nitrate-N leaching losses from lysimeters receiving no urine were between 10 and 11 kg NO₃⁻-N ha⁻¹.

KEYWORDS:

• Beta vulgaris L
• cow urine
• forage crop
• nitrate leaching
• winter grazing

Introduction

In New Zealand, particularly in the South Island, it is common practice to graze pregnant non-lactating dairy cows on high-yielding forage crops during the June–July winter period. This is to minimise treading damage to pasture and soils on the milking platform and to conserve pasture for the following lactation. However, as nitrate (NO₃⁻) leaching losses in grazed systems primarily occur beneath animal urine deposits (Cameron et al. 2013), winter forage grazing can contribute a disproportionately large fraction of whole-farm N leaching losses (Chrystal et al. 2012) because of the high stocking rates and the time of year (i.e. winter, when minimal amounts of N are taken up by vegetation and soils are regularly draining; Edwards et al. 2014b). Simulation modelling by Monaghan et al. (2007) and Chrystal et al. (2012), using the OVERSEER nutrient budgets model (Wheeler et al. 2003), indicated that winter forage crop systems in southern New Zealand could leach 55–60 kg N ha⁻¹ per year into waterways. Furthermore, in situ paddock-scale leaching loss measurements from these systems show N losses can range from 52 to 173 kg N ha⁻¹ y⁻¹ (Shepherd et al. 2012; Smith et al. 2012; Monaghan et al. 2013).

The main winter forage crops grown in the South Island of New Zealand are kale (Brassica oleracea var. acephala L.) and swede (Brassica napobrassica). However, in recent years there has been a significant increase in fodder beet (Beta vulgaris L.) for dairy winter grazing as well, with estimations of 15,000 ha grown in 2014 compared with an estimated 100 ha in 2006 (Gibbs 2014). Given the high yielding nature of fodder beet, and its rapid expansion for agricultural use in recent years, there is concern about additional environmental pressures (especially N leaching losses), with animals stocked at higher densities on fodder beet than on other winter forage crops such as kale (Edwards et al. 2014b).

While recent studies have quantified NO₃⁻-N leaching losses from other winter forage crop systems (e.g. Shepherd et al. 2012; Smith et al. 2012; Hill et al. 2014; Malcolm et al. 2014), until now little attention has been given to winter-grazed fodder beet. The objective of this study was therefore to quantify NO₃⁻-N leaching losses from urine and non-urine areas of a fodder beet winter-grazing system using large intact soil monolith lysimeters.

Materials and methods

Site description and lysimeter collection

The soil type was a free-draining Balmoral/Lismore stony silt loam (Mottled Argillic Pallic Soi,; Hewitt 2010; Udic Ustochrept, Soil Survey Staff 1998) located at the Lincoln University Ashley Dene farm near Lincoln, Canterbury (43°39′ S, 172°20′ E; 17 m above sea level). The site had previously been in permanent pasture for approximately 20 years. Soil cores (n = 50; 0–7.5 cm depth × 2.5 cm diameter) were taken from the site and analysed for soil fertility status prior to the commencement of the trial (Table 1). All nutrients were within acceptable ranges.

Table 1. Initial soil fertility status of the lysimeter collection site prior to basal fertiliser applications in 2012. All units are in ‘MAF quick-test units’ unless specified (Mountier et al. 1966).

Soil property	Value	Methodology
pH	5.9	Water/soil ratio of 2:1 (v/v)
Olsen P (mg L^−1)	29	Olsen et al. (1954)
Sulphate S (mg kg^−1)	21	Searle (1988)
Cation exchange capacity (cmolc kg^−1)	14	Hesse (1971)
Total N (%)	0.43	Keeney & Bremner (1966)
Exchangeable Ca	8	Blackmore et al. (1972)
Exchangeable Mg	10	Blackmore et al. (1972)
Exchangeable K	14	Blackmore et al. (1972)
Exchangeable Na	5	Blackmore et al. (1972)

Undisturbed soil monolith lysimeters (50 cm diameter × 70 cm depth) were collected using the methods described in Cameron et al. (1992). This involved placing a metal cylindrical drum on the soil surface, carefully digging the soil from around the drum, and pushing the drum down with minimal pressure by small increments. The drums were dug down until the soil surface was 5 mm below the top of the casing (this prevented runoff in or out of the lysimeter during the trial). The soil monoliths were then cut off at the base using a hydraulically operated cutting plate, the latter of which was subsequently secured to the base of the lysimeter drum. Petroleum jelly was injected into the annular gap that was intentionally created during the sampling between the soil core and the drum to prevent preferential edge-flow. The lysimeters were then installed into an outdoor trench facility at Lincoln University, Canterbury, New Zealand, with the soil surface of the lysimeters level with the surface of the surrounding field. The space outside the lysimeters was backfilled with soil to the same level as the surface of the lysimeters and plastic tubing was connected to the base of each lysimeter, which fed drainage water into 10 L collection vessels. A set of eight new lysimeters were collected in each year of the trial.

Crop and soil management

Before treatment applications

In December of 2011 and 2012, pasture inside the lysimeters was sprayed with Roundup (glyphosate 360). When senescence had occurred, the top 5 cm depth of soil in each lysimeter was lightly cultivated by hand. ‘Rivage’ fodder beet plants that were c. 8–10 weeks old were taken from an established crop at the Lincoln University Ashley Dene farm (sown in October/November of 2011 and 2012, respectively) and transplanted into the lysimeters at three plants per lysimeter. This was equivalent to a sowing rate of 15 plants m⁻² compared with a recommended commercial rate of 10–12 plants m⁻². The higher rate was used to ensure an adequate number of plants survived the transplanting process.

All lysimeters received urea fertiliser in split applications (25 kg N ha⁻¹) up to a total of 50 kg N ha⁻¹ between February and April in both years. Fertiliser applications were followed with 10 mm of irrigation water, to minimise the risk of volatilisation losses. The irrigation water was applied using TeeJet FL-5VC spray nozzles mounted directly over the top of each lysimeter. In March of 2012 and 2013, superphosphate (9% phosphate, 12% sulphur) was applied at the rate of 750 kg ha⁻¹ to ensure adequate amounts of phosphate and sulphur were maintained.

On 26 June 2012 and 14 June 2013, respectively, all fodder beet plants (bulb and green material together) were pulled and removed from the lysimeters to simulate winter grazing. The soil surface inside the lysimeters was then heavily trampled using a manually operated trampling device as described in Malcolm et al. (2014).

After treatment applications

On 26 October 2012 and 16 September 2013, respectively, weeds inside the lysimeters were sprayed with Roundup (glyphosate 360). Lysimeters were then lightly cultivated and sown with perennial ryegrass seed (Lolium multiflorum, ‘Extreme AR37’) at 25 kg ha⁻¹ to simulate local farm practice. Thereafter, pasture inside each lysimeter was cut to a height of c. 5 cm and removed every 3–4 weeks. Irrigation was applied when necessary to replace evapotranspiration losses and ensure optimal conditions for plant growth.

Cow urine collection and standardisation

Fresh urine was collected directly from pregnant non-lactating dairy cows grazing fodder beet crops on the Lincoln University Ashley Dene farm on 25 June 2012 and 13 June 2013, respectively, using the methods outlined in Miller et al. (2012). The urine collected in 2012 was standardised to produce a concentration of 3 g N L⁻¹ using urea and glycine at a 9:1 ratio, while the urine in 2013 was standardised to 2.5 g N L⁻¹. The urinary N concentrations used were derived from animal urinary spot measurements taken from herds of c. 30 non-lactating dairy cows grazing fodder beet between 1000 h and 1400 h during winter grazing (sampled in both years). Urinary N concentrations measured were c. 3.0 (± 1.5) g N L⁻¹ on average in 2012. In 2013, concentrations ranged between 1.9 and 2.7 g N L⁻¹. The standardised concentration of 2.5 g N L⁻¹ used in 2013 was based on the previous year’s data combined with the measurements made in 2013.

Treatments

In both years the trial was arranged in a completely randomised design, with two cow urine treatments (with and without) replicated four times, giving a total of eight lysimeters per year. Following simulated grazing on both 26 June 2012 and 14 June 2013 as described above, 2 L of the urine collected and standardised the previous day were uniformly applied to each lysimeter to simulate urine deposition during winter grazing (Hogg 1981; Malcolm et al. 2014). This represented urinary N loading rates equivalent to 300 and 250 kg N ha⁻¹, respectively.

Winter rainfall was supplemented to the 75th percentile (if required) of total monthly rainfall (calculated from the 24-year period between 1975 and 1999) through a computer-controlled rainfall simulation/irrigation system (Fraser et al. 1994). This was fully automated and is described in detail in Malcolm (2013).

Measurements

Drainage water from the lysimeters was collected once per week, or when the volume of drainage reached c. 2 L. Total drainage volume was measured and subsamples were analysed for NO₃⁻-N and ammonium-N (NH₄⁺-N) concentration by flow injection analysis (FIA) (Gal et al. 2004; Tecator Inc.). Total NO₃⁻-N leaching losses were calculated from NO₃⁻-N concentrations in the drainage water collected from each lysimeter and the volume of drainage water. Average leaching losses were then calculated using values from the four replicates.

Measurements commenced following urine applications in June (official trial start date) and were carried out for 9–10 months in each of the 2 years (dependent on the length of time for NO₃⁻-N concentrations in the drainage water to return to background values).

Statistical analysis

Differences in peak NO₃⁻-N concentration and total NO₃⁻-N and NH₄⁺-N leaching losses in each season were subjected to analysis of variance (ANOVA) using GenStat (14th edition, Lawes Agricultural Trust), with urine application rate as a fixed effect. Datasets from each year were analysed independently and were log-transformed where necessary to ensure homogeneity of residual errors.

Results

Air temperature, water inputs and drainage

Daily mean air temperatures during both experimental periods were similar to the long-term district means (1971–2000) during the winter months (data not shown).

Daily rainfall, cumulative rainfall + irrigation and cumulative supplementary winter rainfall information for the trial period in both years are presented in Figure 1. A total of 1171 mm of water input was measured by the end of the 2012 trial period (Figure 1A). A total of 177 mm of irrigation was applied as simulated winter rainfall during the months of June–September 2012 before normal irrigation patterns commenced. Simulated winter rainfall application rates ranged from 1–44 mm d⁻¹. During the 2013 trial, lysimeters received 1017 mm of water (Figure 1B). From the time of urine application until 31 September 2013, 193 mm (6–28 mm d⁻¹) was applied as simulated rainfall.

Figure 1. Daily rainfall, cumulative rainfall + irrigation (i.e. all water inputs) and cumulative supplementary winter rainfall (June–September) following cow urine treatment applications to lysimeters. A, 2012; B, 2013.

On average, c. 425 and 350 mm of drainage water was collected from the lysimeters in the 2012 and 2013 trials, respectively (Figure 2). By the end of the spring period in both years (30 November), c. 80%–90% of the total drainage water was recorded.

Figure 2. Nitrate-N concentrations (mg N L^-1) in drainage water following cow urine treatment applications to lysimeters. A, 2012; B, 2013.

Nitrate-N concentrations in drainage water

The application of urine to the lysimeters in both years had a highly significant (P < 0.001) effect on peak NO₃⁻-N concentrations compared with those measured in the control, where no urine was applied (Figure 2). Nitrate-N concentrations in the control treatments peaked at c. 5–7 mg N L⁻¹, while the urine treatments reached peaks of 43 mg N L⁻¹ in early October 2012 (Figure 2A) and 82 mg N L⁻¹ in late August 2013 (Figure 2B). Peak mean NO₃⁻-N concentration was observed after c. 270 mm of drainage water had occured in 2012 and after c. 100 mm of drainage water in 2013.

Total mineral-N leaching losses

Approximately 92%–98% of total mineral-N (NO₃⁻-N + NH₄⁺-N) leaching losses were in NO₃⁻-N form. Nitrate-N leaching losses were strongly affected by urine application in both years (P < 0.001; Table 2). The total NO₃⁻-N leaching loss from the urine treatment in 2012 was equivalent to 64 kg N ha⁻¹, which was 53 kg N ha⁻¹ more than the loss measured beneath the control treatment where no urine was applied. The total amount of NO₃⁻-N leached from the urine-treated lysimeters in 2013 was 79 kg N ha⁻¹. The amount of NO₃⁻-N leaching loss was equivalent to 21% and 32% of the N applied in the urine in 2012 and 2013, respectively (Table 2). By 30 November in each year, > 95% of the total NO₃⁻-N losses measured in drainage had occurred.

Table 2. Total nitrate-N (NO₃⁻-N) leached (means obtained by back-transformation of log10 transformed data) from lysimeters treated with cow urine at 0, 300 (2012) and 250 kg N ha⁻¹ (2013).

Urinary N rate (kg N ha^−1)	Total amount of NO3^−-N leached (kg N ha^−1)	Amount of NO3^−-N leached as % of urinary N applied
2012
0	10.3	N/A
300	63.5	21.2
P value	< 0.001	–
LSR (5%)	2.1	–
2013
0	11.1	N/A
250	78.7	31.5
P value	< 0.001	–
LSR (5%)	1.6	–

LSR, least significant ratio.

Residual d.f. 7.

Discussion

Leaching losses from lysimeters

As expected, there was a significantly greater amount of NO₃⁻-N leached from the urine-treated lysimeters than from the nil-urine control treatment lysimeters. Further, as a proportion of the amount of N applied to the lysimeters in the urine, an equivalent of 21%–32% was leached as NO₃⁻-N. In the study of Malcolm et al. (2014) where N leaching losses beneath winter kale were reported, there was a higher proportional leaching loss, equivalent to 43% and 54%, when urine was applied at the higher rates of 500 and 700 kg N ha⁻¹, respectively. This is similar to trends observed in grazed pasture systems (Cameron et al. 2013), suggesting an exponential relationship between urine-N loading rate and N leaching losses under winter forage crop systems.

Nitrate-N concentration curves varied markedly between years, with a pronounced peak earlier in 2013, but a smaller and less distinct peak later in 2012 (Figure 2). A number of high rainfall events shortly after urine application in 2013 was likely responsible for the rapid leaching shown by the NO₃⁻-N peak in 2013 (Figure 1). Peak NO₃⁻-N concentrations in the drainage water were also low compared with values reported in other studies (McDowell & Houlbrooke 2009; Malcolm et al. 2014) and this can be attributed to the lower urinary N rates of 300 and 250 kg N ha⁻¹ used in our study. Interestingly, the total NO₃⁻-N leaching loss in 2013 was apparently higher than in 2012, despite the lower amount of urinary N applied. This was probably due to the large rainfall events shortly after urine application in June 2013 (Figure 1), coupled with warmer temperatures in 2013 during July and August that may have enhanced both mineralisation and nitrification under the 250 kg N ha⁻¹ treatment, and cooler temperatures during spring in 2012 under the 300 kg N ha⁻¹ treatment.

Nitrate-N leaching losses from urine areas of simulated winter-grazed fodder beet in our study were lower than those of some previous work investigating kale (e.g. Hill et al. 2014; Malcolm et al. 2014), most likely because of the low urinary N concentrations measured from cows grazing fodder beet. Urinary N concentrations are largely driven by N intake (Beukes et al. 2014) and, given that crude protein concentrations of fodder beet are generally lower than those of other forages, particularly pasture (Chakwizira et al. 2014; Rugoho et al. 2014), N intakes are consequently also lower (i.e. less than 300 g N cow⁻¹ day⁻¹; Edwards et al. 2014a). Low rates of N excretion from cows grazing fodder beet were also recently reported in a study by Edwards et al. (2014a), with average urine concentrations ranging from 1.9 to 2.3 mg N L⁻¹ from cows grazing fodder beet during two winters in Canterbury. For validation purposes, more detailed work is required to compare urinary N concentrations from cows offered different winter forage crop diets. This should also include diurnal effects (i.e. how concentrations may vary over a 24-h period).

Conclusions

The application of cow urine to lysimeters resulted in a mean total leaching loss of 64 and 79 kg NO₃⁻-N ha⁻¹ in the 2012 and 2013 seasons, respectively. By and large the losses measured were low compared with previous studies, driven mainly by the low urinary N concentration. These results can provide a first basis for future assessments of NO₃⁻ leaching losses from winter grazing of fodder beet at different scales.

Acknowledgements

We thank the following people for technical support: Trevor Hendry, Roger Atkinson, Nigel Beale and Neil Smith for technical assistance; Jiao Zhang and Barry Anderson, Lincoln University, for sample analysis; and Esther Meenken, The New Zealand Institute for Plant & Food Research Limited, for advice on the statistical analysis of the data.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research work was conducted through the Pastoral 21 Programme (C10X1117), jointly funded by the Ministry of Business, Innovation and Employment, DairyNZ Inc., Fonterra, Beef + Lamb New Zealand and the Dairy Companies Association of New Zealand.