Inland waters are an essential component of the global carbon cycle as they are very active sites for carbon transformation processes. Much of this carbon is transformed into the greenhouse gas carbon dioxide (CO₂) and emitted into the atmosphere. The biogeochemical and hydrological mechanisms driving CO₂ concentrations in inland waters are manifold. Although some of them have been studied in detail, there are still knowledge gaps regarding the relative importance of the different CO₂-driving mechanisms, both on a spatial and a temporal scale. The main aim of this thesis was to fill some of the knowledge gaps by studying long- and short-term effects of enhanced dissolved organic carbon (DOC) concentrations on surface water partial pressure of CO₂ (pCO₂) as well as to investigate both internal (i.e., within the water body) and external (i.e., catchment) drivers of pCO₂ in inland waters. Based on analyses of long-term data from more than 300 boreal lakes and streams and on results from two mesocosm experiments as well as a detailed catchment study, one of the main results of the thesis was that DOC concentrations were, on a temporal scale, generally uncoupled to pCO₂. Indeed, additions of allochthonous DOC to lake water could result in increased pCO₂ in waters but not as originally expected by stimulation of bacterial activity but instead by light driven suppression of primary production, at least in mesotrophic waters. Changes in the carbonate system was also found to be a main driver for surface water pCO₂. Finally, also external processes such as groundwater inputs contributed substantially to variations of surface water pCO₂. In a detailed study on carbon in groundwater, pCO₂ in groundwater was found to decrease with soil depth and correlated negatively with pH, which increased with soil depth. Conclusively, this thesis show that pCO₂ does not follow the trends of increased DOC in boreal surface waters but instead correlates with changes in primary production and shifts in the carbonate system. Additionally, the dominating mechanisms driving pCO₂ clearly differ between lakes and streams. Consequently, simulations of future CO₂ dynamics and emissions from inland waters cannot rely on DOC concentrations as a pCO₂ predictor, but rather need to incorporate several pCO₂ driving mechanisms, and consider the difference between lakes and streams.