Inland waters play an active role in the global carbon cycle and emit large volumes of the greenhouse gases (GHGs), methane (CH4 ) and carbon dioxide (CO2 ). A considerable body of research has improved emissions estimates from lakes, reservoirs and rivers but recent attention has been drawn to the importance of small, artificial waterbodies as poorly quantified but potentially important emission hotspots. Of particular interest are emissions from drainage ditches and constructed ponds. These waterbody types are prevalent in many landscapes and their cumulative surface areas can be substantial. Furthermore, GHG emissions from constructed waterbodies are anthropogenic in origin and form part of national emissions reporting, whereas emissions from natural waterbodies do not (according to Intergovernmental Panel on Climate Change guidelines). Here, we present GHG data from two complementary studies covering a range of land uses. In the first, we measured emissions from nine ponds and seven ditches over a full year. Annual emissions varied considerably: 0.1-44.3 g CH4 m-2 year-1 and -36-4421 g CO2 m-2 year-1 . In the second, we measured GHG concentrations in 96 ponds and 64 ditches across seven countries, covering subtropical, temperate and sub-arctic biomes. When CH4 emissions were converted to CO2 equivalents, 93% of waterbodies were GHG sources. In both studies, GHGs were positively related to nutrient status (C, N, P), and pond GHG concentrations were highest in smallest waterbodies. Ditch and pond emissions were larger per unit area when compared to equivalent natural systems (streams, natural ponds). We show that GHG emissions from natural systems should not be used as proxies for those from artificial waterbodies, and that artificial waterbodies have the potential to make a substantial but largely unquantified contribution to emissions from the Agriculture, Forestry and Other Land Use sector, and the global carbon cycle.

The transport of dissolved organic matter (DOM) across the land-ocean-aquatic-continuum (LOAC), from freshwater to the ocean, is an important yet poorly understood component of the global carbon budget. Exploring and quantifying this flux is a significant challenge given the complexities of DOM cycling across these contrasting environments. We developed a new model, UniDOM, that unifies concepts, state variables and parameterisations of DOM turnover across the LOAC. Terrigenous DOM is divided into two pools, T 1 (strongly-UV-absorbing) and T 2 (non- or weakly-UV-absorbing), that exhibit contrasting responses to microbial consumption, photooxidation and flocculation. Data are presented to show that these pools are amenable to routine measurement based on specific UV absorbance (SUVA). In addition, an autochtonous DOM pool is defined to account for aquatic DOM production. A novel aspect of UniDOM is that rates of photooxidation and microbial turnover are parameterised as an inverse function of DOM age. Model results, which indicate that ~ 5% of the DOM originating in streams may penetrate into the open ocean, are sensitive to this parameterisation, as well as rates assigned to turnover of freshly-produced DOM. The predicted contribution of flocculation to DOM turnover is remarkably low, although a mechanistic representation of this process in UniDOM was considered unachievable because of the complexities involved. Our work highlights the need for ongoing research into the mechanistic understanding and rates of photooxidation, microbial consumption and flocculation of DOM across the different environments of the LOAC, along with the development of models based on unified concepts and parameterisations.