Carbon dioxide emissions to the atmosphere from inland waters-streams, rivers, lakes and reservoirs-are nearly equivalent to ocean and land sinks globally. Inland waters can be an important source of methane and nitrous oxide emissions as well, but emissions are poorly quantified, especially in Africa. Here we report dissolved carbon dioxide, methane and nitrous oxide concentrations from 12 rivers in sub-Saharan Africa, including seasonally resolved sampling at 39 sites, acquired between 2006 and 2014. Fluxes were calculated from published gas transfer velocities, and upscaled to the area of all sub-Saharan African rivers using available spatial data sets. Carbon dioxide-equivalent emissions from river channels alone were about 0.4 Pg carbon per year, equivalent to two-thirds of the overall net carbon land sink previously reported for Africa. Including emissions from wetlands of the Congo river increases the total carbon dioxide-equivalent greenhouse-gas emissions to about 0.9 Pg carbon per year, equivalent to about one quarter of the global ocean and terrestrial combined carbon sink. Riverine carbon dioxide and methane emissions increase with wetland extent and upland biomass. We therefore suggest that future changes in wetland and upland cover could strongly a ect greenhouse-gas emissions from African inland waters.C limate predictions necessitate a full and robust account of natural and anthropogenic greenhouse-gas (GHG) fluxes, especially for CO 2 (refs 1-3), CH 4 (ref. 4) and N 2 O (ref. 5), which together accounted for 94% of the anthropogenic global radiative forcing by well-mixed GHGs in 2011 relative to 1750 (ref. 6). Inland waters (streams, rivers, lakes and reservoirs) are increasingly recognized as important sources of GHGs to the atmosphere, with global CO 2 and CH 4 emissions estimated at 2.1 PgC yr −1 (ref.3) and 0.7 PgC yr −1 (CO 2 -equivalents; CO 2 e) (ref. 4) (1 Pg = 10 15 g), respectively. Considering that the oceanic and land carbon (C) sinks correspond to ∼1.5 and ∼2.0 PgC yr −1 (ref. 7), respectively, the GHG flux from inland waters is significant in the global C budget.In a recent global compilation of inland CO 2 data 3 , <20 data points (out of 6,708, that is, <0.3%) represented African inland waters (with the exception of South Africa, which has been densely sampled), even though they account for ∼12% of both global freshwater discharge 8 and riverine surface area 3 , and include some of the largest rivers and lakes in the world. Equally for the global CH 4 database, there is a strong under-representation of tropical inland waters, whereby a recent synthesis 4 resorted to extrapolating CH 4 fluxes from temperate rivers.The prevailing large uncertainty involved in GHG flux estimates for inland waters, essentially due to the paucity of available data, is coupled to a poor understanding of underlying processes, both of which preclude gauging of future fluxes in response to human pressures. In particular, there is a need to further understand the link between inland water GHG fluxes and ...

The role played by river networks in regional and global carbon (C) budgets is receiving increasing attention. Despite the potential of radiocarbon measurements (Δ 14 C) to elucidate sources and cycling of different riverine C pools, there remain large regions for which no data are available and no comprehensive attempts to synthesize the available information and examine global patterns in the 14 C content of different riverine C pools. Here we present new 14 C data on particulate and dissolved organic C (POC and DOC) from six river basins in tropical and subtropical Africa and compiled >1400 literature Δ 14 C data and ancillary parameters from rivers globally. Our analysis reveals a consistent pattern whereby POC is progressively older in systems carrying higher sediment loads, coinciding with a lower organic carbon content. At the global scale, this pattern leads to a proposed global median Δ 14 C signature of À203‰, corresponding to an age of~1800 years B.P. For DOC exported to the coastal zone, we predict a modern (decadal) age (Δ 14 C = +22 to +46‰), and paired data sets confirm that riverine DOC is generally more recent in origin than POC-in contrast to the situation in ocean environments. Weathering regimes complicate the interpretation of 14 C ages of dissolved inorganic carbon, but the available data favor the hypothesis that in most cases, more recent organic C is preferentially mineralized.

Abstract. Inland waters have been recognized as a significant source of carbon dioxide (CO 2 ) to the atmosphere at the global scale. Fluxes of CO 2 between aquatic systems and the atmosphere are calculated from the gas transfer velocity and the water-air gradient of the partial pressure of CO 2 (pCO 2 ). Currently, direct measurements of water pCO 2 remain scarce in freshwaters, and most published pCO 2 data are calculated from temperature, pH and total alkalinity (TA). Here, we compare calculated (pH and TA) and measured (equilibrator and headspace) water pCO 2 in a large array of temperate and tropical freshwaters. The 761 data points cover a wide range of values for TA (0 to 14 200 µmol L −1 ), pH (3.94 to 9.17), measured pCO 2 (36 to 23 000 ppmv), and dissolved organic carbon (DOC) (29 to 3970 µmol L −1 ). Calculated pCO 2 were > 10 % higher than measured pCO 2 in 60 % of the samples (with a median overestimation of calculated pCO 2 compared to measured pCO 2 of 2560 ppmv) and were > 100 % higher in the 25 % most organic-rich and acidic samples (with a median overestimation of 9080 ppmv). We suggest these large overestimations of calculated pCO 2 with respect to measured pCO 2 are due to the combination of two cumulative effects: (1) a more significant contribution of organic acids anions to TA in waters with low carbonate alkalinity and high DOC concentrations; (2) a lower buffering capacity of the carbonate system at low pH, which increases the sensitivity of calculated pCO 2 to TA in acidic and organicrich waters. No empirical relationship could be derived from our data set in order to correct calculated pCO 2 for this bias.Owing to the widespread distribution of acidic, organic-rich freshwaters, we conclude that regional and global estimates of CO 2 outgassing from freshwaters based on pH and TA data only are most likely overestimated, although the magnitude of the overestimation needs further quantitative analysis. Direct measurements of pCO 2 are recommended in inland waters in general, and in particular in acidic, poorly buffered freshwaters.

Abstract. Inland waters have been recognized as a~significant source of carbon dioxide (CO2) to the atmosphere at the global scale. Fluxes of CO2 between aquatic systems and the atmosphere are calculated from the gas transfer velocity and the water-air gradient of the partial pressure of CO2 (pCO2). Nowadays, direct measurements of water pCO2 remain scarce in freshwaters and most published pCO2 data are calculated from temperature, pH and total alkalinity (TA). Here, we compare calculated (pH and TA) and measured (Equilibrator and headspace) water pCO2 in a large array of temperate and tropical freshwaters. The 761 data points cover a wide range of values for TA (0 to 14.2 mmol L−1), pH (3.94 to 9.17), measured pCO2 (36 to 23 000 ppmv), and dissolved organic carbon (DOC) (29 to 3970 μmol L−1). Calculated pCO2 were > 10% higher than measured pCO2 in 60% of the samples (with a median overestimation of calculated pCO2 compared to measured pCO2 of 2560 ppmv) and were > 100% higher in the 25% most organic-rich and acidic samples (with a median overestimation of 9080 ppmv). We suggest these large overestimations of calculated pCO2 with respect to measured pCO2 are due to the combination of two cumulative effects: (1) a more significant contribution of organic acids anions to TA in waters with low carbonate alkalinity and high DOC concentrations; (2) a lower buffering capacity of the carbonate system at low pH, that increases the sensitivity of calculated pCO2 to TA in acidic and organic-rich waters. We recommend that regional studies on pCO2 should not be based on pH and TA data only, and that direct measurements of pCO2 should become the primary method in inland waters in general, and in particular in acidic, poorly buffered, freshwaters.