Abstract Stream metabolism provides insight into the functional processes that regulate trophic status, biomass, and nutrient cycling in streams. Comparative (spatial) analyses of streams generally do not account for shifting at‐site controls on stream metabolism over time and, thus, may not identify the primary controls on cross‐site differences in trophic status, biomass, and nutrient cycling over longer time scales. The spatial component included assessing environmental factors controlling stream metabolism (daily values) during a summer low‐flow period (August) at 17 wadable stream sites in four regions of the United States. Explanatory variables for the spatial analysis included discrete nutrients and physical parameters. The temporal component included assessing 12 of the 17 sites with more than 90 days of stream metabolism values, and continuous nitrate, photosynthetically active radiation (PAR), turbidity, maximum water temperature (°C) and stream stage (water surface elevation). Spatial analysis at the 17 sites during August (31 days) found that average gross primary production (GPP) ranged from 0.1 to 4 g O2 m−2 day−1 and ecosystem respiration (ER) from 0.37 to 6.4 g O2 m−2 day−1. Sites were primarily heterotrophic; however, some sites varied from autotrophic to heterotrophic daily. Multiple regression indicated that GPP was a function of orthophosphate, percent canopy cover, and the number of days since a high flow (R2 = 0.51), whereas ER was primarily a function of GPP (R2 = 0.40). Temporal patterns in daily stream metabolism were evaluated using at‐site autoregressive models at 12 streams with relatively complete, continuous records. PAR, nitrate and maximum water temperature were the dominant variables influencing GPP, and water depth and nitrate were the most common explanatory variables for ER models. The autoregressive components (i.e., GPP or ER for the previous day) had a strong influence on both GPP and ER except after spates, indicating that biomass may be the dominant control on metabolism rather than variability in environmental factors during stable‐flow periods. We found two dominant temporal metabolism regimes: a seasonal pattern with elevated spring metabolism followed by low stable summer metabolism and a temporally disturbed pattern which tended to occur in streams with frequent spates. These patterns were not indicated by differences in metabolism among streams during stable low‐flow conditions. Because streams are generally heterotrophic with only limited periods when production may be high, comparative analyses must estimate stream metabolism over a time scale representing at‐site spatial variability and account for controls on periods of high production to adequately characterise differences between streams.