|dc.description.abstract||Globally, agriculture contributes 10-12% to anthropogenic greenhouse gas (GHG) emissions. Consequently, mitigation of agricultural GHGs has taken on increased importance, particularly in countries like New Zealand where agriculture accounts for almost half of national emissions. Sequestration of atmospheric CO₂ in the soil by altering land management practices has been identified as a potential mitigation option for anthropogenic GHG emissions. However, implementing management practices as a mitigation option first requires an understanding of their effect on soil C stocks. Often the effects of cropping and grassland management on soil C stocks are studied individually, but these practices are linked when the resulting crop is supplied to grazing animals as supplemental feed. This is important for the New Zealand dairy industry, which has been traditionally pasture-based, but increasingly is supported by supplemental feed to compensate for periods of low pasture growth and to boost productivity. To understand the impact of supplemental feed use on soil C stocks and, therefore, on any mitigation potential, both the production of the feed and its use need to be considered together. The overarching aim of this thesis was to experimentally determine the impacts of supplemental feed production and use on soil C by using the net ecosystem carbon balance (NECB) methodology to quantify changes in ecosystem carbon (C) stocks (assumed synonymous to the change in soil C). A secondary aim of this thesis was to advance NECB methodology in complex grazed pasture systems, primarily through examination of the scale at which measurements were made. Improved methodology and understanding are needed to allow a greater number of management practices to be tested.
Conceptually, importation of supplemental feed and its embodied C can lead to an increase in ecosystem C because consumption of supplemental feed C by the animals results in additional excreta deposition on the pasture during grazing that can be stored as soil C. This hypothesis was tested by determining the NECB for three years on a dairy farm where imported supplemental feed accounted for >40% of the cows’ diet. A positive NECB (indicating a gain of ecosystem C) was calculated for all three years, but consideration of uncertainties resulted in only one year having a definitive gain of C. The three-year average NECB was 71 ± 77 g C m⁻² y⁻¹ (mean ± uncertainty) and was not considered different from zero. Theoretical calculations based on the imported quantity of supplemental feed C (average 526 g C m⁻² y⁻¹) coupled with the digestibility of the feed and manure retention rates suggest gains of around 25 g C m⁻² y⁻¹ could be expected. The results of this study were of the same order of magnitude to what was expected from modelling and manure C retention literature, and although experimentally a gain in C associated with a large import of supplemental feed could not be definitively concluded, the results confirmed that large gains of ecosystem C are unlikely.
A broad range of supplemental feed is used within the New Zealand dairy industry including grazed and harvested feeds, with maize harvested for silage being one of the more common. Internationally, sites where maize cropping with full biomass harvest occurs have been identified as a large source of C, but these studies tend to be from long-term cropping systems. Within New Zealand dairy farm systems, maize silage is often grown as part of the pasture renewal process and, consequently, findings from studies within long-term cropping systems may not apply. In this study, NECBs were calculated for a system where a long-term pasture site was converted to maize silage cropping for two years before a return to permanent pasture. To isolate the C balance of the maize crop alone, NECBs were calculated for the period of maize crop establishment through to seedling emergence of the subsequent sward (~190 days). The Year 1 maize crop NECB was –850 g C m⁻² (a loss of C), while the Year 2 maize crop lost a further –415 g C m⁻². Concurrent grazed pasture NECBs from the same farm were 11 g C m⁻² and –115 g C m⁻² over the same two periods. Above-ground biomass production was around three times greater from the maize crop than adjacent pastures, with more than 90% of this production exported from the site, compared to around 60% net export of the pasture biomass after accounting for returned excreta. The hypothesis that a large loss of ecosystem C could be expected from maize silage cropping for supplement feed was supported. Future research to determine whether the return to permanent pasture results in recovery of previously lost C are required to understand the long-term impacts of periodic cropping for supplemental feed production.
Consideration of the effect that all types of supplemental feed production have on ecosystem C stocks was beyond the scope of this thesis, but conclusions can be drawn on systems which use maize silage. Dairy farms which import supplemental feed (maize silage or other) are likely to see small increases (<50 g C m⁻² y⁻¹) in the ecosystem C stocks regardless of the quantity imported, while the production site would be expected to have large losses when producing maize silage. Results from this thesis suggest that where production and use occur within the same dairy farm system a net loss would be expected, and if averaged across the entire farm would be in the order of –40 g C m⁻² y⁻¹. Losses during maize silage production may be reduced by minimising the time that soil is bare during establishment, and although not tested, possibly by decreasing tillage intensity. Moreover, if ecosystem C losses during production are recovered longer-term when returned to grazed pasture (i.e. in the several years following cropping), on-farm production of periodically cropped maize silage may lead to small, but consistent gains in soil C and provide the potential for GHG mitigation. A key unresolved question is the rate of C recovery following a return to permanent pasture relative to the cropping return period.
While determining the effect of supplemental feed on ecosystem C, the opportunity also arose to investigate two aspects of NECB measurement scale. Firstly, NECBs were compared when calculated with an ecosystem boundary equivalent to (i) the paddocks included within the eddy covariance (EC) flux footprint (NECBFootprint), and (ii) the farm boundary (NECBFarm). Both calculated NECBs were similar (NECBFootprint was 56 ± 77 g C m⁻² y⁻¹ and NECBFarm was 71 ± 77 g C m⁻² y⁻¹) and the selection of the best boundary definition was dependent on the quality of the available data with NECBFarm considered best in this study. Furthermore, components contributing to the NECB differ with system boundary location and, therefore, can influence interpretation. When choosing a system boundary, the assumption that the measured CO₂ exchange is representative of the entire area within the defined boundary needs to be cautiously considered. The second methodology investigation calculated paddock-specific NECBs for two adjacent paddocks with a single EC system located between them. Provided regular EC data are available from both paddocks (i.e. due to regularly changing wind directions), paddock-specific NECBs can be calculated. Advantages of this approach include eliminating inherent management heterogeneity (e.g. asynchronous grazing), and the ability to allow for treatment comparisons or provide replication while minimising spatial variability and potentially reducing equipment requirements. Key disadvantages were a reduction in data coverage (from 49.1% for the full footprint to 25.9% and 15.7% for each paddock), an increase in uncertainty (by about 25%), and the need for prior assessment of site suitability (i.e. the need for regular wind from both paddocks). Comparisons of NECBs from adjacent rotationally grazed paddocks identified large inter-annual and between-paddock variability, with the latter often due to subtle management differences despite the same overall management regimes. Finally, due to the spatial and temporal variability, several measurement years would be needed to (i) determine the true trajectory of ecosystem C balances and (ii) determine similarity or differences between the two paddocks.||