A nested sampling network on the Colorado (CR) and Missouri Rivers (MR) provided data to assess impacts of large-scale reservoir systems and climate on carbon export. The Load Estimator (LOADEST) model was used to estimate both dissolved inorganic and organic carbon (DIC and DOC) fluxes for a total of 22 sites along the main stems of the CR and MR. Both the upper CR and MR DIC and DOC fluxes increased longitudinally, but the lower CR fluxes decreased while the lower MRs continued to increase. We examined multiple factors through space and time that help explain these flux patterns. Seasonal variability in precipitation and temperature, along with site-level concentration versus discharge relationships proved to be significant factors explaining much of the difference among sites located below reservoirs as compared to sites located in more free-flowing segments of the river. The characterization of variability in carbon exports over space and time provides a basis for understanding carbon cycling and transport within river basins affected by large reservoir systems, particular in arid-to semi-arid ecosystems.