Diffuse agricultural pollution is a major contributor to poor water quality in many parts of the world. Consequently agri-environment policy promotes the use of riparian buffer strips and/or denitrifying wetlands to intercept and remove diffuse NO3--N pollution. However, these methods have the potential to cause ‘pollution swapping’: the exchange of one form of pollution as a result of measures implemented to reduce another. Thus the benefits of intercepting NO3--N could be offset by enhanced emissions of the potent greenhouse gases, nitrous oxide (N2O) and methane (CH4), from buffer strips and wetlands. This research aimed to: (1) quantify the direct N2O emissions from an irrigated buffer strip (IBS), using nitrate-rich agricultural drainage water, compared to a non-irrigated control (CP); (2) improve the understanding of N2O production and consumption within soils using controlled soil monolith experiments; (3) assess the effectiveness of a small (60 m2) instream wetland at intercepting and removing diffuse NO3--N pollution, and quantify pollution swapping in the form of CH4 and N2O emissions; (4) assess the production of CH4 and N2O within the sediment, and their emissions as well as inorganic-N concentrations in the overlying water column in response to temperature and turbulence, using intact wetland sediment and membrane inlet mass spectrometry (MIMS). The research focused on mitigating diffuse NO3--N pollution from grazed pasture at a farm in north-east England. Annual N2O-N emissions from the IBS and CP were not statistically different (P > 0.05): 509 and 263 g N2O-N ha-1, respectively, in 2007 and 375 and 500 g N2O-N ha- 1 year-1, respectively, in 2008. Irrigation of the IBS increased spatial variability in flux and generated hotspots of denitrification compared to the CP. However, these changes were short-lived. Direct N2O emission factors (EF1) calculated using the available NO3--loading data (September 2007 - December 2008) for the IBS were lower (c.0.1%) than those calculated for the CP assuming N input from biological N fixation only (<1.9%). Soil monolith experiments under a variety of irrigation and NO3--N loading regimes confirmed low direct and indirect (of dissolved N2O-N in leachate) emissions (<3.1 and <2.3% of applied NO3--N emitted as N2O-N, respectively), similar to the IPCC default emission factors. However, N loss in leachate was high, up to 82% of added NO3--N with concentrations reaching 24 mg NO3--N L-1. Therefore even though no pollution swapping occurred the high leachate losses indicate irrigation of buffer strips are not effective mitigation methods. Monitoring for 2 years of the instream wetland that received median NO3--N concentrations of c. 6 mg N L-1, but up to c. 20 mg N L-1, showed it to be ineffective at intercepting diffuse NO3- pollution: likely a result of the relatively high discharge and short water residence time, as well as the direct input of NO3--N to the wetland from secondary sources: field drains and/or overland flow. The wetland was a net source of NH4+-N in both 2007 and 2008, and a net sink of NO3--N in 2007 only. Annual wetland CH4 and N2O emissions were 713 and 237 mg CH4 m-2 year-1, and 3.5 and 1.9 mg N2O-N m-2 year-1, for 2007 and 2008, respectively and were highly variable between seasons. N pollution swapping was minimal from either direct or indirect emissions, but CH4 emissions were found to be of greater importance at a net cost of ~ £600 ha-1 over the study period (2007 to 2008), compared to N2O emissions (~ £60 ha-1) and low NO3--N interception savings (~ £24 ha-1). Incubation experiments suggest that spatially variable microsites of nitrifying, denitrifying or methanogenic activity and CH4 oxidation occur within the wetland sediment. Therefore off-line, larger wetland systems offer the best prospects of enhanced NO3--N interception and potentially reduced CH4 emissions by maintaining shallow water depths (increased CH4 oxidation) and long residence times (increased opportunity for denitrification), within the wetland or wetland cells.