Separation of the net ecosystem exchange of CO2 (F) into its component fluxes of net photosynthesis (FA) and nonfoliar respiration (FR) is important in understanding the physical and environmental controls on these fluxes, and how these fluxes may respond to environmental change. In this paper, we evaluate a partitioning method based on a combination of stable isotopes of CO2 and Bayesian optimization in the context of partitioning methods based on regressions with environmental variables. We combined high‐resolution measurements of stable carbon isotopes of CO2, ecosystem fluxes, and meteorological variables with a Bayesian parameter optimization approach to estimate FA and FR in a subalpine forest in Colorado, United States, over the course of 104 days during summer 2003. Results were generally in agreement with the independent environmental regression methods of Reichstein et al. (2005a) and Yi et al. (2004). Half‐hourly posterior parameter estimates of FA and FR derived from the Bayesian/isotopic method showed a strong diurnal pattern in both, consistent with established gross photosynthesis (GEE) and total ecosystem respiration (TER) relationships. Isotope‐derived FA was functionally dependent on light, but FR exhibited the expected temperature dependence only when the prior estimates for FR were temperature‐based. Examination of the posterior correlation matrix revealed that the available data were insufficient to independently resolve all the Bayesian‐estimated parameters in our model. This could be due to a small isotopic disequilibrium (equation image) between FA and FR, poor characterization of whole‐canopy photosynthetic discrimination or the isotopic flux (isoflux, analogous to net ecosystem exchange of 13CO2). The positive sign of equation image indicates that FA was more enriched in 13C than FR. Possible reasons for this are discussed in the context of recent literature.