Much of the climate change impact of agriculture is due to the addition of nitrogen (N) to soil to increase crop yields. The production of mineral N fertiliser from atmospheric N2 causes carbon dioxide (CO2) emissions, and once added to soil, fertiliser N is easily transformed to other forms and lost as leached nitrate (NO3-), ammonia gas (NH3) and the greenhouse gas nitrous oxide (N2O). An alternative to fertiliser N is the use of green manures, plants grown on agricultural land, which after incorporation into soil provide N to the following crop. N fixing green manures gain atmospheric N by a symbiosis with microbes, therefore adding N without CO2 emissions, and providing beneficial organic matter to soil. Counteracting these advantages is a lower land use efficiency due to the space required for green manure cultivation. Green manure additions can result in carbon (C) sequestration, however soil C dynamics are complex and soil C can also be lost by a stimulation of respiration known as priming.

N pollution from soil is generally higher when mineral N concentrations are high, which can result from poor synchrony of N supply with crop demand. The incorporation of green manures offers limited scope for effective targeting of N. This synchrony of N supply with demand could be improved by, instead of incorporating green manures, harvesting the plant material and adding it to soil to meet crop demand. Negating the requirement for incorporation, allows use of a wider range of species including shrubs and trees. Growing N fixers suited to lower quality land e.g. wet or exposed areas, could increase land use efficiency by reducing the demand for prime agricultural land. We chose three such N fixing species and refer to these as Perrenial Mobile Green Manures (PMGMs): Alnus glutinosa (Alder), Gunnera manicata (Gunnera), and Ulex europaeus (Gorse). This thesis investigated N provision to a crop, and potential for N pollution by these PMGMs, compared to the conventional green manure, Trifolium pratense (red clover) and ammonium nitrate (NH4NO3) fertiliser, in a one year pot experiment, and field experiments over two seasons. A six week incubation investigated soil C dynamics by use of 14C isotopes.

In both pot and field experiments, the PMGMs supplied N to crops at a slower rate than clover or NH4NO3 but by the end of the pot experiment resulted in equal or more biomass and N uptake than from clover. Potential for N pollution from PMGMs was considerably lower than from clover or NH4NO3, with mean NO3- concentrations in the soil solution of the pot experiment reaching only 25 mg N L-1 compared to over 130 mg N L-1 from clover and NH4NO3 additions. Emissions of N2O from PMGMs were considerably lower than those from clover and NH4NO3, in the pot trail, with applied N lost as N2O-N from gunnera and alder being 0.34 %, and 0.61 %, respectively compared to 5.3 % from clover. N2O emissions from all treatments in the field experiment were low which is likely to be due to dry weather conditions. Data from the incubation study indicated that after one year, gunnera and alder additions could result in a net loss of C due to priming. These predictions, however, do not consider factors of cropping soil e.g. roots, and meso and macro fauna and require further investigation. We conclude that PMGMs could improve N use efficiency and reduce NO3- leaching and N2O emissions compared to incorporation of clover, and have potential for a more favourable C balance than NH4NO3. Thus, PMGMs have strong potential for inclusion into a more sustainable agricultural landscape.