Significance of wetlands in the agricultural landscape as sources of nitrous oxide emissions
Authors: Jim Cooke, Kit Rutherford, John Wilcock, Fleur Matheson
Executive summary
Download the full report in PDF format [PDF 4.8MB]
Agriculture contributes half of New Zealand’s greenhouse gas emissions and is the major source of methane (88%) and nitrous oxide (96%). While the sources of methane are well known and undertood, there is much less certainty about nitrous oxide. As is the case internationally, indirect emissions (of which leaching and runoff make up about 75%) account for a disproportionate share of the uncertainty in the estimates. If we are to manage nitrous oxide emissions, it is important that we gain a better understanding of the emission sources. Riparian wetlands within agricultural catchments are a likely source of nitrous oxide since they are natural conduits for nitrate-rich leachate and are known to have a high capacity for denitrification. Since retention and creation of wetlands within agricultural landscapes are promoted as a means of reducing the nitrate load to surface waters, it is important that we have an increased understanding of the role of wetlands in nitrous oxide emissions relative to that of the pasture catchment.
Because there have been no New Zealand studies that have explicitly measured nitrous oxide emissions from riparian wetlands, we abstracted information from international literature in order to obtain a range of emission rates from wetland systems, and to glean the principal processes and environmental conditions leading to nitrous oxide production. We used the understanding gained from this review to develop conceptual and empirical models. We used an empirical models to put some bounds on the likely flux arising from riparian wetlands in New Zealand pasture catchments, and as a basis for planning field measurements/experiments to fill knowledge gaps.
Nitrous oxide emission rates in wetland and riparian wetland studies overseas range from 0 (or net sink) to ~ 100 kg N ha-1 y-1 . In wetlands receiving runoff from agricultural fields, rates are commonly in the range 0 - 45 kg N ha-1 y-1 with the high end of the range associated with high nitrate input and removal rates. Zero or low emission rates are associated with permanent innundation and low nitrate inputs. The range of emission rates in wetlands is within the range of rates obtained in terrestrial ecosystems, and high nitrogen input is a common factor in high emission rates measured in both systems. Far fewer studies have been carried out on wetlands than pasture and cropping systems. In addition there is evidence of distinct spatial trends of nitrous oxide emissions in riparian wetlands compared with a more random pattern regulated by water filled pore spaces in terrestrial systems. Riparian wetland studies that do not take account of this spatial pattern may calculate a low average nitrous oxide emission overall, but mask very high rates occurring within small areas. The identification of these small areas of riparian wetland may be key to minimising emissions from this source.
There are a number of microbially-mediated biochemical pathways (nitrification, denitrification, anammox, dissimilatory nitrate reduction to ammonia) that can produce nitrous oxide as intermediate products. Nitrification is frequently invoked as a significant (or even the main) pathway in aerobic soils, but in wetlands, where the supply of oxygen is limited, the evidence suggests that denitrification is mostly responsible. Contemporary microbiological research, however, cautions against ascribing nitrous oxide flux to a single process or pathway because the microoganisms are physiologically defined, and are widespread amongst various taxa that vary in their responses to environmental conditions. Therefore, from a mangement viewpoint it is more important to understand the environmental factors leading to high rates of nitrous oxide emissions rather than the microbial processes per se.
Isolating individual factors influencing nitrous oxide emissions in a riparian wetland environment is difficult although insights have come from multivariate analysis. Further knowledge has been derived from detailed measurements at points along transects through riparian zones; in conjunction with laboratory studies where single environmental variables can be isolated. From a comparison of the ‘state of the knowledge’ with our understanding of New Zealand riparian wetlands we concluded that the main variables affecting nitrous oxide production are nitrate and oxygen concentration. Whilst pH and available carbon have been shown to be important in certain circumstances, our view is that they are of lesser importance in the New Zealand context. Maximum nitrous oxide (as a proportion of other products) occurs at oxygen concentrations that are sub-optimal for both nitrification and denitrification and some studies have shown this also corresponds to maximum nitrous oxide production. Because oxygen measurement at low levels in field situations is notoriously difficult, surrogates such as redox potential or water filled pore space are commonly used. In riparian wetlands, redox potential is a more appropriate measure and peak nitrous oxide production has been shown to occur over a very narrow range of +120-250 mV. This range of redox potential has also been shown to correspond to the zone where nitrate is predicted to be reduced under equilibrium conditions. Field studies have also shown that wetlands receiving high external loads of nitrate from the surrounding catchment have much higher nitrous oxide fluxes than those dependent on autochthonous (nitrification-induced) nitrate supply from water table fluctuations. The further reduction of nitrous oxide to dinitrogen appears to be regulated by nitrate supply.
Nitrous oxide emissions from riparian wetlands depend not only on the environmental conditions influencing nitrous oxide production but also the physical factors affecting the escape of nitrous oxide molecules to the atmosphere. Overlying water decreases the diffusion of nitrous oxide by ~ 4 orders of magnitude, which partly explains why permanently flooded wetlands (such as constructed wetlands receiving wastewater) have relatively low nitrous oxide emissions. Riparian wetland plants can provide the gaseous diffuson pathway that circumvents the barrier induced by overlying water, but this appears to be quite plant-specific. In general some aerobic interface appears to be necessary for significant nitrous oxide emission; whether it is through wetting and drying processes (as occurs seasonally in many New Zealand pastoral riparian wetlands) or at the upland/wetland interface (where pasture-derived nitrate leachate first contacts the saturated organic riparian soil). There is evidence that higher fluxes of nitrous oxide are associated with such breaks in the landscape and New Zealand studies have shown that this interface is associated with extremely high in situ denitrification rates (although nitrous oxide itself was not measured).
A number of authors have stated that predicting the ratio of nitrous oxide:dinitrogen gas is key to understanding nitrous oxide emissions from riparian wetlands. Recent Dutch studies have disputed that simple relationships exist between this ratio and environmental variables (pH, Eh, oxygen, available carbon etc). Nevertheless, faced with few other quantifiable predictors of nitrous oxide flux we envisaged that this ratio would be a pragmatic way of putting some upper bounds upon emissions from New Zealand riparian wetlands. We used a ‘hole in the pipe’ model to conceptualise the losses of nitrous oxide from a ‘stream’ of nitrate entering one end of a pipe, and dinitrogen gas (indicative of complete nitrate reduction) exiting the other end. However, we conclude that although the ‘hole in the pipe’ model is a useful conceptual model, it is currently of limited value for calculating nitrous oxide fluxes in New Zealand wetlands, because; (a) there are few simultaneous measurements of nitrous oxide emission and denitrification in wetlands with high denitrification rates and (b) it ignores the findings of spatial pattern of denitrification and nitrous oxide emission rates.
Complex models such as the DNDC model (even the wetland-DNDC model) are not suitable (in the first instance) for estimating nitrous oxide flux from wetlands in New Zealand agricultural landscapes because of a lack of input data. We have therefore constructed a simpler empirical model incorporating nitrate flux and known denitrification rates from pasture, wetland, and headwater stream environments. The model showed that increasing the area of riparian wetlands is likely to increase nitrous oxide emissions from the catchment if (i) the denitrification rate of headwater streams is low compared with the flux of N from the catchment, and (ii) if the N yield from the catchment exceeds the denitrification rate of the wetlands. Under scenarios where the denitrification potential of downstream rivers, floodplains and estuaries is comparable with the cumulative flux of N exported from upstream catchments then total emissions will remain constant; all that changes is the location where they occur.
There is no suitable inventory of wetlands in pasture catchments we can use to estimate the significance of notrous oxide emissions from this source, however one study in the Lake Taupo catchment showed that wetlands comprised > 5% of the total farmed area. Whilst our empirical model enables us to put some approximate bounds on nitrous oxide emssions from wetlands in New Zealand agricultural landscapes, we have no way of validating the model. There is, we believe, good justification for obtaining systematic measurements of nitrous oxide emissions coupled with environmental variables (such as nitrate flux, Eh, available carbon) across pasture-wetland transects. This will enable us to not only confirm or deny the predictions made in our model, but also make a significant international contribution to the understanding of nitrous oxide emissions and how to manage them.
Completing the additional objectives proposed under this programme will allow improved understanding of the spatial variability of nitrous oxide emissions across riparian wetlands, and, together with estimates of their area, allow a robust estimate to be made of the total nitrous oxide emissions wetlands within pasture catchments.
Contact for Enquiries
Sustainable Land Management and Climate Change
MAF
Pastoral House
25 The Terrace
PO Box 2526, Wellington
Tel: 0800 CLIMATE (254 628)
Contact this person
