Abstract Soils annually emit between 6.8 and 7.9 Gt CO 2 equivalents, mainly as CH 4 from intact peatlands and from rice agriculture; as N 2 O from unmanaged and managed soils; and as CO 2 from land-use change. Methane emissions attributable to other wetlands add another 1.6–3.8 Gt CO 2 equivalents. From a global standpoint, N 2 O from unmanaged soils and CH 4 from peatlands and other wetlands make soils naturally net greenhouse gas emitters. In addition, the storage of carbon in soils and the fluxes of CH 4 and N 2 O have been changed by anthropogenic effects towards emission rates 52 to 72% above those under natural conditions before the dawn of intensive agriculture and land-use change. Land-use changes on mineral soils induced most of the recorded losses of soil organic matter (SOM), but there is evidence that proper agricultural management of soil resources is able to recover some of these losses and to maintain soil functions. However, the discrepancy between so-called ‘sequestration potentials’ and the measures already adopted is amazingly large. Globally, only about 5% of the cropped areas is managed according to practices such as no tillage or organic farming. The contribution of soil loss by erosion, desertification and sealing to global oxidative SOM losses is uncertain; however, in the case of soil erosion, it is considered to be a major factor in global SOM decline. Mitigation options calculated for SOM restoration, reduced CH 4 and N 2 O emissions are able to alleviate mean annual emissions by 1.2 to 2.9 Gt CO 2 equivalents, mainly as a result of carbon sequestration, which is the most efficient measure for the next few decades. In the longer term, however, the large potential for reducing CH 4 and N 2 O emissions outweigh the finite capacity of soils to recover C. Integrated assessment of net greenhouse-gas fluxes is key for evaluating management practices aimed at reducing overall emissions. From the viewpoint of climate change and taking into consideration the mean fluxes of CO 2 , CH 4 and N 2 O, peatland protection is more favourable than peatland cultivation in the long term. The most important gaps in our understanding appear to be with regard to estimating fluxes along with soil erosion and desertification processes, in the extent of peatland cultivation; the role of black carbon formation, natural ‘background’ sequestration rates of undisturbed soils; and the net response of soils, particularly in cold regions, to global warming. With regard to the societal perception of soil contributing to the global cycling of greenhouse gases, it is important to emphasize that significant proportions of the emissions are inevitably linked to intensive agriculture.