Reducing nitrogen loss from cow urine patches: Strategies for pasture-based dairy systems : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

In dairy grazing systems, intensive farming practices with high nutrient inputs contribute to environmental pollution with nitrogen (N) loss from farmland, a key contributor to water quality degradation. The New Zealand dairy industry aims to reduce N leaching by 40-60% by 2030. Achieving this target may require transitioning from moderate to lower input systems and implementing mitigation strategies focused on managing high N sources, such as cow urine patches, to improve N use efficiency and reduce leaching risk. While cow urine patches present a primary source of N loss from pasture-based dairy farms, there is limited information on how the relationship between urine volume per urination event and urine patch area affect the N leaching risk. The aim of the thesis is to investigate this relationship and how grazing management strategies influence urine N load and plant N uptake following a regrowth period to mitigate N leaching risk. Additionally, it aims to develop a tool for predicting the impact of management strategies on N loss from urine patches. A key finding of the literature review was that successful N mitigation requires tactical and operational knowledge of implementing mitigation strategies, which has slowed adoption due to uncertainty in implementation. Additionally, the review also identified mitigations centred around the transfer of N, specifically relating to the urine patch, urine N load, or plant N uptake, to improve N use efficiency within a farm system. The review identified a lack of decision support tools for farmers to tailor N mitigations to individual circumstances. One finding was the opportunity to provide tactical and operational tools, such as a urine patch model, to understand urine distribution variation and develop management strategies that target urine and urine patch characteristics to reduce leaching risk. Using a modelling approach (Chapter 3), the aim was to determine if reducing N input alone could achieve the required N leaching reduction with minimal impact on farm operating profit, or if additional mitigation practices targeting N transfer from high N sources, such as cow urine patches, are needed to improve N use efficiency and reduce the risk of leaching. A two-year dairy farmlet study was conducted comparing two farm systems: moderate stocking rate (MSR, 3.9 cows/ha); and lower stocking rate (LSR, 2.9 cows/ha); with a benchmark, high-performing commercial demonstration farm (Lincoln University Dairy Farm [LUDF]; 3.4 cows/ha). Milk yield, pasture production, and quality data were collected, and modelled in FARMAX and OverseerFM to estimate the financial and environmental performance of each farm system. The LSR system produced the best environmental outcome across the two years (2018/19 and 2019/20), leaching an estimated 31% less N compared with MSR and LUDF, but at the cost of profitability. The average, annual milk solid production per ha was 28% less for LSR relative to MSR and LUDF. Correspondingly, the average annual operating profit per ha was 35% less for LSR compared with LUDF. A low N input system reduces production and operating profit to the extent that these mitigation strategies may not be adopted. Therefore, the next step was to target sources of high N input, such as cow urine patch areas, and improve N use efficiency from these areas to reduce the risk of leaching. In Chapter 4, we investigated the effects of grazed pasture canopy characteristics (using plant species and height) on urine patch area. The experiment was conducted during autumn in Canterbury, New Zealand. Warm water was used to simulate urine events, ranging in volume from 1 to 8 L, onto either partially (Lenient, 6-15 cm) or fully (Hard, <6 cm) grazed perennial ryegrass and white clover or pure plantain pastures. A thermal digital camera and imaging software were used to calculate the wetted area of each urine event. When a mid-range volume (4 L) was poured onto lenient grazed pastures, plantain had a greater wetted area than perennial ryegrass/white clover pasture (0.30 m2 ± 0.11 m2 and 0.16 m2 ± 0.08 m2, respectively; mean ± standard deviation). However, the wetted area was similar for plantain and perennial ryegrass/white clover pasture under hard grazing (0.36 m2 ± 0.16 m2 and 0.31 m2 ± 0.13 m2, respectively). Irrespective of pasture treatments and grazing intensity, the relationship between water volume and wetted area was curvilinear, with no significant increase in wetted area for simulated urine events greater than 4 L. Our results indicated that both pasture treatments and grazing intensity (i.e., residual pasture canopy) affect urine patch area, which could have potential implications for the urine N load per urine patch. A limitation of this study was that sward surface height was measured at the paddock level rather than for each urine event simulated. The first field experiment (Chapters 5 and 6) was split into two chapters to evaluate two objectives. The first objective was to investigate the effect of pre-graze pasture mass and time of pasture allocation, on total soil N levels from urine patches and subsequent N recovery from pasture following a regrowth period. The experimental design was a 2 x 2 x 2 factorial arrangement of treatments replicated twice within two experimental runs. The first factor corresponded to two levels of pasture mass using moderate to high mass (2226 or 2662 kg DM/ha respectively) to create lower herbage crude protein content in the high mass treatment. The second factor corresponded to time of allocation of pasture using morning or afternoon allocation. The third factor corresponded to time of urine deposition using anticipated peak (dawn) vs nadir (mid-morning) urinary N concentration and urine N load. A two-way interaction indicated larger urine patch areas for moderate mass at peak times (P < 0.05). As well as a three-way interaction which showed that moderate mass allocated in the afternoon resulted in larger urine patch areas (P < 0.05). Total soil N content was 6% higher (P < 0.05) for moderate mass but did not differ (P > 0.05) between peak and nadir times. High mass treatments showed better regrowth in both urine and non-urine patches. For non-urine patches, the herbage N yield was 18% lower for the moderate compared to the high mass pastures (P < 0.05). From this, we concluded that managing pasture mass can influence pasture recovery and nitrogen utilisation, with potential implications for reducing N leaching risk from dairy farms. Chapter 6 aimed to investigate the relationship between volume per urination, urine patch area, and compressed pasture height. The objective of this chapter was to understand if pasture management practices can affect urine patch area and subsequent urine N load. Cows were fitted with acoustic urine sensors (AgResearch, New Zealand) on their hind legs. The sensors captured the timing, duration, volume per urination, and frequency of urination. Thermal digital imaging with an automated algorithm (AgResearch) was used to estimate urine patch area. Our findings revealed that compressed pasture height did not affect the urine volume and patch area relationship. The relationship was also weaker compared to the findings in Chapter 4. The results showed the importance of accurate urine volume measurement and the impact that measurement errors with different methodologies could have on the relationship with urine patch area. Further research with accurate measurements of volume per urination and corresponding patch area is needed to develop tools to model urine and urine patch characteristics and N leaching risk accurately before they can inform management strategies. The second field experiment (Chapter 7) was carried out to collect additional data on the urine volume and patch area relationship. This study investigated the impact of pasture treatments and height on this relationship through simulated urination events. Conducted at Lincoln University Research Dairy Farm, the experiment utilised a 2 x 8 factorial design, comparing perennial ryegrass-white clover pasture with Italian ryegrass-plantain-red and white clover pasture across eight volume treatments (1-8 L). The study found no significant difference in urine patch area between pasture treatments. A curvilinear relationship between urine volume and patch area was identified (similar to that in Chapter 4), with a quadratic model (Urine patch area = 0.129 + 0.056v – 0.003v²) explaining 54.8% of the variance. Additionally, sward surface height significantly influenced urine patch area, with taller pastures reducing urine patch area, especially at higher urine volumes. This interaction was modelled (Urine patch area = 0.1259 + 0.09278v – 0.004742v² – 0.002363vh + 0.0001293v²h), explaining 70% of the variance. In summary, effective strategies must combine reduced N inputs with mitigation practices targeting N transfer from high N sources, such as cow urine patches, to improve N use efficiency and reduce leaching risk. This thesis has shown that altering pasture mass or the timing of pasture allocation does not significantly influence urine patch area; instead, it was primarily affected by the time of urine deposition. Additionally, these grazing management practices do not enhance plant N uptake from urine patches. Management practices aimed at reducing urine N load at peak times might be more effective at reducing N leaching risk than grazing management. However, grazing management can increase N transfer during the regrowth phase from small urine events. By managing sward surface height when smaller urine events fall within the linear part of the volume and patch area relationship, the urine patch area can be increased at the time of deposition, improving urine N load distribution and plant N uptake. Ultimately, the urine patch model indicated that reducing daily N intake and increasing daily urine volume throughout the year can reduce N leaching by up to 14.5%.