Codenitrification under ruminant urine patch conditions: microbial contributions, substrates and nitrogen flux kinetics : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

A large fraction of anthropogenic nitrous oxide (N2O) emissions can be traced back to grazed grasslands, where N2O is mainly emitted from livestock-urine affected soils (so called ‘urine patches’). Since N2O is a potent greenhouse gas and an ozone-depleting substance, research has focused on its formation pathways inside agricultural soils in order to mitigate agricultural N2O emissions. In recent years, evidence was found for the significance of N2O-formation pathways, creating so called “hybrid N2O” where a soil derived nitrogen (N) source is co-metabolised with an applied N source (e.g. urine-urea or mineral N fertilizers). This reaction(s) is referred to as ‘codenitrification’ and may, besides denitrification and nitrification, be responsible for up to 95% of the emitted gaseous N-compounds. First reports of the potential for microorganisms to produce hybrid N2O and/or hybrid dinitrogen (N2) were almost last 100 years ago, however, the current state of knowledge about codenitrification remains sparse. For example, it still remains unclear what the relative contributions of different microbial groups are, within a pasture soil context, or what the co-metabolites might be dominant in such a process as codenitrification. The number of studies relating to codenitrification performed in vitro by far outnumber the studies dealing with soils. Thus soil mesocosm experiments are required in order to gain more insight into codenitrification reactions responsible for N2O emissions from urine patches. Filling this important knowledge gap is an essential step for furthering N2O emission mitigation strategies. Two experiments were carried out in 2015-2016 (year one) and 2017 (year two) using freshly collected soil from the nearby Lincoln University Research Dairy Farm. The first experiment focused on the fungal and bacterial contributions to hybrid N2O emissions following a simulated bovine urine event. Fungal and bacterial inhibitors were used in order to inhibit the different microbial groups individually or collectively, and while utilizing applied 15N-labelled urea at a rate of 1000 kg N ha-1. The stable isotope approach allowed N transformations to be traced, as performed over time by the microbial groups. It was demonstrated that under elevated soil pH conditions, codenitrification generated hybrid N2O and N2, with codenitrified N2O accounting for > 30% of the total N2O flux. The inhibition of fungi lead to a reduction of ≥ 42% of the codenitrification derived N2O flux while the bacterial inhibition did not cause a significant reduction in codenitrification. Despite there being mainly bacterial driven ammonia (NH4+) oxidation, the results demonstrated that soil fungi were the main organisms undertaking codenitrification. The second experiment was designed to identify some of the co-metabolized compound(s) utilized for hybrid N2O formation. Soil mesocosms were prepared and (non 15N-labelled) urea applied (equalling 500 kg N ha-1). Eight days afterwards, when codenitrification was theoretically able to occur, four different potential 15N-labelled co-metabolites (so called ‘nucleophiles’) were applied to the soil mesocosms to measure their relative contribution to hybrid N2O formation. Glycine and phenylalanine were studied, as they are two amino acids occurring within the soil organic matter (SOM) matrix, and these were complemented with NH4+ and hydroxylamine (NH2OH), each applied at a rate of 20 µg N g-1 soil. The same inhibitors as used in the first experiment were used to identify the microbial groups using the nucleophiles. This time, only codenitrified N2O was measured. All applied nucleophiles were observed to contribute to hybrid N2O formation, indicating that soil microbes are capable of using a wide range of possible N compounds for codenitrification. The most reactive nucleophile, in biotically and abiotically mediated reactions, was NH2OH. Finally, a modelling approach was used to investigate the underlying N transformations rates and N pool developments in pasture soil subsequent to a simulated urine event. Data from a previous experiment with known codenitrification fluxes were used to run a model, suitable for modelling the gross N transformation rates, prior to the N2O formation. The modelled output matched well with the measured data and revealed significant changes in the labile and recalcitrant fraction of soil organic matter and their related transformation rates subsequent to the urea application. Evidence was found for the NH4+, NO3-, labile and recalcitrant N pool to be involved in codenitrification reactions. Especially the labile N pool was assumed to provide possible nucleophiles, consisting not only of easy degradable N compounds but also of N loosely bound to organic carbon components due to reactions of NH3 with dissolved organic carbon. Furthermore, an increasing enrichment of SOM with 15N of up to 23 atm% 15N after 1519 h (63 d) and at the same time, 43% and 29% (wet and dry soil, respectively) of the applied N were still stored in clay minerals. This clearly affects the following N transformations further and, compared with the previously detected codenitrification fluxes, indicates that in addition with free NO3-, a high soil moisture content and increased SOM are favouring codenitrification. This body of work shows that codenitrification is closely linked with the activity of fungi and that a wide range of organic and mineral N compounds can be utilized as nucleophiles.