Effect of urinary nitrogen, dairy shed effluent and nitrogen fertiliser on nitrate leaching from a pasture soil

Abstract

Environmental concerns over nitrate-N contamination of surface and groundwater have led to a number of studies on nitrate-N leaching from grasslands. Most studies of nitrate-N leaching loss from grassland have focused solely on single nitrogen (N) sources (e.g. fertiliser. urine, dung or effluent). Under grazed pasture, however, more than one N source is usually returned to the system either directly or indirectly. Therefore it is important to improve our knowledge and understanding of the concentrations, quantities and transport mechanisms of nitrate-N from grazed pasture receiving N in single or combined inputs. In this study two field experiments were carried out using undisturbed Templeton fine sandy loam (Udic-Ustochrept, coarse loamy, mixed, mesic) soil lysimeters (500 mm diameter by 700 mm deep) with ryegrass (Lolium perenne L) - white clover (Trifolium repens L) pasture. The effect of the application of cow urine, dairy shed effluent (DSE) and N fertiliser either individually or collectively on the concentrations and quantities of nitrate-N leached were determined from the two year lysimeter experiment. The role of macropore flow on nitrate-N leaching was determined in a second lysimeter experiment. The results from the two year experiment were also used to test the accuracy and the utility of the LEACHN model to predict nitrate-N leaching losses.

In the two year lysimeter experiment, cow urine was applied to the lysimeters, at rates of 0 and 1000 kg N ha⁻¹ in May 1996 and 1997. In the first year (May 1996) bromide was added as a conservative tracer prior to urine application. In the second year (May 1997) urine was labelled with ¹⁵N isotope as a tracer in two urine treatments in order to determine a complete mass balance from urine. Urea and DSE were applied to urine and non-urine treated lysimeters at rates of 0, 200 and 400 kg N ha⁻¹ in 4 split equal applications in May, August and November and February. Natural rainfall was supplemented with simulated rainfall in the winter and spring (May to October) to achieve the 75th percentile of the winter/spring rainfall records in the region. Flood irrigation was applied 6 times during summer/autumn (November to April) at 100 mm per application which is a typical practice used by dairy farmers in the region. The lysimeters were managed by cutting the pasture to 40-70 mm heights at typical grazing intervals. Drainage water was collected and analysed for nitrate-N, nitrite-N, ammonium-N, bromide and ¹⁵N.

Over the period of the experiment (May 1996 to April 1998), the dominant form of N leached was nitrate-N. The annual average nitrate-N concentrations in the drainage from the lysimeters which received urine alone were between 26-38 mg L⁻¹. Corresponding values for urine + DSE (33-51 mg L⁻¹) and urine + urea (40-69 mg L⁻¹) treatments were higher than the urine alone treatment. These nitrate-N concentrations were significantly higher than those treatments which did not receive urine (1-5 mg L⁻¹). The amounts of nitrate-N leached from the urine alone treatment were found to range between 77 and 124 kg ha⁻¹. The combination of urine and DSE resulted in total nitrate-N leaching losses of between 90 and 168 kg ha⁻¹. The combination of urine and urea fertiliser resulted in the highest nitrate-N losses, ranging from 156 to 190 kg ha⁻¹. Total losses from non-urine treatments were significantly lower (1-18 kg N ha⁻¹) than those from the urine treatments.

The total mineral N leaching loss from the urine alone treatment represented 10% of the total N applied over the two year period. When urine (1000 kg N ha⁻¹) was applied in combination with DSE (400 kg N ha⁻¹), the mineral N leaching loss was equivalent to 7% of the total N applied. Application of urine (1000 kg N ha⁻¹) together with urea (400 kg N ha⁻¹), resulted in a mineral N leaching loss of 12% of the total N applied. The relatively lower loss from the urine and DSE combined treatment was probably because of the different chemical forms of N applied and the supply of carbon in that treatment enhancing denitrification and immobilisation. Corresponding values for non-urine treatments were significantly lower (1-3%).

It is important to recognise that on average only about 25% of the area of a grazed dairy paddock receives urine per year. Therefore the measured losses under the urine treated lysimeters will be diluted by leachate from non-urine areas of the paddock (e.g. 75% of the paddock). The field scale leaching losses were thus calculated by taking into account the dilution effect of drainage water from non-urine patch areas of the paddock using the following equation:

Np = N₁ x 0.25 + N₂ x 0.75

Where Np is the calculated annual average concentration of nitrate-N under the paddock (mg L⁻¹) or annual N leaching loss (kg ha⁻¹) under the paddock, N₁ is the annual average concentration or amount of nitrate-N leached from the urine, or urine plus DSE or urea treatments, which were measured from lysimeters, and N₂ is the annual average concentration or amount of nitrate-N leached from the control, DSE and/or urea treatments alone, measured from the lysimeters.

The calculated annual paddock nitrate-N concentrations were 7-17 mg L⁻¹ and annual paddock losses were 20-60 kg ha⁻¹.

The effect of macropore flow on urine-N leaching was determined by conducting experiments using a tension infiltrometer to impose suctions of 0.5 kPa and 0 kPa on top of a lysimeter that was treated with urine. The 0.5 kPa suction prevented soil pores greater than 600 µm diameter from conducting water and solutes, while the 0 kPa suction allowed conduction under 'field saturated' conditions. It was found that pores greater than 600 µm diameter transmitted about 98% of the total N leached below 700 mm depth and that the main form of N leached from the urine under 0 kPa suction was ammonium-N. The maximum peak concentration of ammonium-N (11 mg L⁻¹) accounted for 10.5% of the total N applied at 0 kPa suction. This was significantly higher than the peak concentration of ammonium-N (1.5 mg L⁻¹) and percentage of the total N leached (0.2%) at 0.5 kPa suction. The peak concentrations of urea-N in the leachate reached 16 mg L⁻¹ at 0 kPa suction, and accounted for 1.6% of the total N applied. No urea-N was detected in the leachate at the 0.5 kPa suction. The concentrations and amounts of nitrate-N leached were very low and did not differ between the two suctions. The forms and amounts of N leached were affected by the macropore flow and the effect this had on the opportunity for N transformation to occur in the soil.

The LEACHN model predicted nitrate-N leaching for cow urine accurately, but was unable to provide an accurate prediction for non-urine treatments. However the utility of the model for management purposes was considered to be limited because of the need for large number of parameter values which could only be obtained through intensive experiments.