The partitioning of evapotranspiration (ET) between plant transpiration (E_t) and direct evaporation (E_d) presents one of the most important and challenging problems for characterizing ecohydrological processes. The exchange of water vapor (q) and CO₂ (c) are closely coupled in ecosystem processes and knowledge of their controls can be gained through joint investigation of q and c. In this study we examine the correlation of water vapor and CO₂ (R_qc) through analyses of high‐frequency time series derived from eddy covariance measurements collected over a suburban grass field in Princeton, NJ during a 2 year period (2011–2013). R_qc at the study site exhibits pronounced seasonal and diurnal cycles, with maximum anticorrelation in June and maximum decorrelation in January. The diurnal cycle of R_qc varies seasonally and is characterized by a near‐symmetric shape with peak anticorrelation around local noon. Wavelet and spectral analyses suggest that q and c are jointly transported for most eddy scales (1–200 m), which is important for ET partitioning methods based on flux variance similarity. The diurnal cycle of the transpiration fraction (ratio of E_t to total ET) exhibits an asymmetric diurnal cycle, especially during the warm season, with peak values occurring in the afternoon. These ET partitioning results give similar diurnal and seasonal patterns compared with numerical simulations from the Noah Land Surface Model using the Jarvis canopy resistance formulation.