There is a common perception that no-till agriculture directly and always translates to lower soil-related greenhouse gas emissions. The reality is much more complicated.
The IPCC guidelines for GHG inventories assume that soil organic carbon reaches a spatially-averaged stable value over time, specific to the soil, climate and land-use/management practices. But the IPCC does not relate soil N2O emissions to the soil carbon sequestration/emission processes, whereas in reality there is a strong connection.
A study of agricultural practices in Eastern Canada (Gregoich, et al) showed that no-till can increase N2O emissions depending on soil density and water content. N2O emissions for clay soil under no-till were 2-3 times that of conventional till, whereas N2O was lower under no-till for loamy/sandy soils. This was further confirmed in another Canadian study by Rochette, et al, which demonstrated that N2O emissions were exceptionally high in heavy-clay soil after initiation of no-till and more than double the N2O under conventional-till.
Another study (Six, et al) analyzed a large number of available data sets and showed that N2O emissions were higher in no-till than conventional-till for the first 10 years after transition, regardless of climate -- but a more humid climate increased the N2O substantially. The greater soil water content following the adoption of no-till stimulates denitrification of the soil organic matter stocks, resulting in higher N2O emissions for many years as well as a directly related nitrogen deficiency in the soil. Gregorich, et al, observe that relatively high N2O emissions from no-till soils would offset part of the mitigation benefit of increased soil carbon storage. Rochette, et al, point especially to the limited potential of no-till to reduce net GHG emissions in fine-textured soils rich in organic matter that are prone to high water content and reduced aeration.
Moreover, adoption of no-till in dry climates leads to an initial loss of soil organic carbon in the first 5-10 years, possibly due to a slower incorporation of surface residues into the soil than in conventional-till (Six, et al). Putting all this together, the net soil-derived global warming potential is higher for no-till than conventional-till in the first 5-10 years after adoption of no-till. After 20 years, no-till has a significantly lower net GWP in humid climates, but any benefit in dry climates appears to be within the range of uncertainty.
Thus, IPCC's soil carbon model (linear increase/decrease in soil carbon and not linked to the N2O fluxes) is clearly optimistic in calculating the effect of switching from conventional-till to no-till. Caution is definitely in order when estimating the GHG emissions benefits of no-till agriculture.
A related issue in soil dynamics is the so-called steady-state or equilibrium. The idea that soil organic carbon eventually reaches equilibrium (on average) for a given set of conditions is widely accepted by many authors (for example: Smith, et al; Phetteplace, et al ). Every major LCA study of food and agricultural products has assumed the soil carbon to be in this equilibrium state.
Among other recent studies, Soussana, et al measured the net biome productivity (NBP, net of all GHG fluxes from soil and biomass) of managed and unmanaged grasslands. They did not detect an equilibrium for the unmanaged sites -- this at first appears to challenge the notion of carbon sink saturation -- but this was because the measured NBP included not just soil carbon but also fluxes related to the increasing biomass. A recent report from the UK Soil Association (based on a broad survey of published studies) accepts that soil carbon is likely to reach a steady-state value for a given set of conditions. The UK report also points out that long-term organic farming has produced 20-28% higher soil carbon levels in various countries -- but it only considers soil carbon content and does not include N2O emissions, so actual GHG reductions are likely to be smaller.
Overall, I have not found any evidence in the literature that contradicts the soil-carbon equilibrium assumption. The idea of equilibrium makes sense mathematically from a dynamical system perspective. The rate at which carbon accumulates in the soil and the ultimate saturation level are site specific. IPCC's linear/monotonic tier 1 model simply provides a first approximation to this. The part that is not in the IPCC model is the N2O emission mechanism as it relates to land-use/management changes -- this is a significant missing piece that needs to be included in any accurate assessment of agriculture-related mitigation of GHG emissions.