Whole-ecosystem metabolism is an important indicator of the role of organic matter, C cycling, and trophic structure in rivers. Ecosystem metabolism is well studied in small streams, but less is known about metabolism in large rivers. We estimated daily whole-ecosystem metabolism over 2 y for 1 site each at the Mississippi and Chattahoochee Rivers in the USA to understand factors influencing temporal patterns of ecosystem metabolism. We estimated rates of gross primary production (GPP), community respiration (CR), and net ecosystem production (NEP) with a curve-fitting approach with publicly available discharge (Q), dissolved O₂, temperature, and photosynthetically active radiation (PAR) data. Models were run for week-long blocks, and power analyses suggested that rates should be established at least once for each 10-wk period throughout the year to characterize annual rates of metabolism accurately in these 2 rivers. We analyzed weekly rates averaged over 10-wk periods with Spearman rank correlation to identify potential drivers and with path analyses to identify interactions among variables driving GPP, CR, and NEP. Both rivers had an overall negative NEP, and the Mississippi River had stronger seasonal trends. In the Mississippi River, CR was strongly positively correlated with Q, which suggests variation in seasonal availability of allochthonous C. In the Chattahoochee, CR was most strongly positively correlated with GPP, whereas GPP was negatively correlated with Q, which suggests that autochthonous processes and water-column light attenuation played important roles in C dynamics. Our results suggest that these large rivers were net heterotrophic at annual time scales but autotrophy can be important seasonally.