Inland waters are significant sources of carbon dioxide to the atmosphere, and estimates of emissions are similar in magnitude to those of the carbon dioxide sequestered by the net terrestrial sink. Currently, methods of estimating carbon dioxide emissions are based on statistical approaches and often do not consider landscape attributes such as human development, agriculture, or the hydrologic connectivity of the stream network. The following research addresses these issues in chapter 1 by developing and validating a reactive transport model at the watershed scale, then in chapter 2 by applying the reactive transport model at the continental scale across US stream networks, and in chapter 3 by determining the impacts of anthropogenic land uses on carbon dioxide fluxes. Here we present the findings of this research, which show that current statistical methods of predicting carbon dioxide emissions from inland waters overestimate by roughly 25% on continental scales and potentially more when used on individual basins. Beyond this, we shed light on the sources of carbon dioxide emissions from inland waters showing the importance of internally produced carbon dioxide, which accounts for ~87% of emissions across the continental United States stream network and groundwater contributions which dominate headwater streams. Finally, we show that anthropogenic impacts account for small changes in carbon dynamics in streams which are still yet to be fully understood.