According to the UK's climate change programme, agriculture accounted for 12% of the UK's greenhouse gas emissions in 1990, or some 24.8 million tonnes of carbon (mtC).
Emissions from land use change stood at some 8.7mtC - partly offset by an estimated 2.9mtC which was removed from the atmosphere and stored in forests, vegetation and soils. One recent study indicated that the organic carbon content of UK topsoils has declined by between 4 and 23% over the last 20 years.
Estimates of both emissions and carbon sequestration from agriculture sinks are notoriously inaccurate. The project for DEFRA, carried out by agricultural consultants ADAS, set out to provide a better understanding of carbon capture in soils - and the potential for increasing it.
In a striking conclusion, ADAS questions the "current focus" on soil sequestration as the primary contribution of agriculture to greenhouse gas abatement. "The potential for genuine carbon sequestration to soil by agricultural management changes is very limited under British conditions," ADAS says. A "far larger contribution" could come from reducing farm energy use and tackling emissions of the powerful greenhouse gases nitrous oxide and methane.
Indeed, the report warns that the benefits of carbon sequestration in soil can be cancelled out by emissions of other greenhouse gases. Examples include the addition to soil of livestock wastes, sewage sludge, domestic refuse and composts.
Some management choices may be finely balanced. Shallower tillage increases soil CO2, and could sequester up to 0.2mtC per year if applied across the UK. Zero tillage could sequester almost four times as much - but the additional benefits are "almost entirely negated" by increased N2O emissions.
However, some changes appear to have clear advantages. Converting land from arable and grassland usage to woodlands would have the biggest effect. A 10% change - which would make Britain's woodland cover similar to the rest of Europe - could sequester an initial 9mtC per year, the report says.
Energy crops such as short rotation willow or miscanthus would sequester about a sixth as much carbon as permanent woodland - but would also displace carbon emissions from fossil fuel combustion (ENDS Report 325, p 29 ). Energy crops should also be several times cheaper to establish and grow, ADAS argues.
Other changes to farming practices would have more modest benefits. Shifting from fertiliser-based dairy systems to clover systems could save up to 0.6mtC per year. Spreading manure on arable land instead of pasture could deliver another 0.2mtC.
Converting arable land to set-aside by extending field margins would have a "considerable impact", ADAS says, delivering savings of up to about 0.8mtC. Incorporating straw residues back into the land could sequester perhaps 0.3mtC per year - but there is already a market for straw in animal bedding, and the carbon benefits of using straw for energy production may be higher.
Overall, ADAS says, a combination of arable measures could save or sequester perhaps 1.3mtC per year across the whole of England. However, soil sequestration would account for only a small fraction of this - the main benefits would come from lower energy use and lower N2O emissions through reduced fertiliser use.
A framework for monitoring changes needs to be established before major land use changes are implemented, ADAS concludes. It calls for more emphasis on measuring changes in agricultural N2O and methane emissions - but questions whether verification of changes in soil carbon levels are cost-effective given their relatively minor importance.