The purpose of this study was to determine the repeatability of methane (CH₄) and carbon dioxide (CO₂) emissions from beef cattle using GreenFeed emissions monitoring (GEM) system and as affected by sampling frequency and measurement periods. Twenty-eight crossbred replacement beef heifers were monitored using the GEM system over 59 d to collect their CH₄ and CO₂ emissions data. Heifers’ feed intake was recorded by eight automated feeding stations. The standardized dry matter intake (SDMI), CH₄ and CO₂ emission and yield (g kg^-1 SDMI) were averaged over 1, 3, 7, and 14 d periods. On average, animals emitted 204.7 g d^-1 (SD = 36 g d^-1) and 6408 g d^-1 (SD = 780 g d^-1) of CH₄ and CO₂, respectively. Between-animal coefficients of variation for all variables decreased with an increasing averaging period (from 1 to 14 d). The coefficient of determination (R²) between CH₄ emission and SDMI was increased from 0.25 to 0.73 as averaging period increased from 1 to 14 d. Similarly, the R² between CO₂ emission and SDMI increased from 0.39 to 0.79 as averaging period increased from 1 to 14 d. It was determined that averaging over 7 to 14 d with minimum of 20 spot samples was needed to produce repeatable and reliable averaged CH₄ and CO₂ emissions and correlated with SDMI.