Effective isolation of tritium (3H) and other contaminants at waste-burial facilities requires improved understanding of transport processes and pathways. Previous studies documented an anomalously widespread (i.e., theoretically unexpected) distribution of 3H (>400 m from burial trenches) in a dry, sub-root-zone gravelly layer (1–2-m depth) adjacent to a low-level radioactive waste (LLRW) burial facility in the Amargosa Desert, Nevada, that closed in 1992. The objectives of this study were to: (i) characterize long-term, spatiotemporal variability of 3H plumes; and (ii) quantify the processes controlling 3H behavior in the sub-root-zone gravelly layer beneath native vegetation adjacent to the facility. Geostatistical methods, spatial moment analyses, and mass flux calculations were applied to a spatiotemporally comprehensive, 10-yr data set (2001–2011). Results showed minimal bulk-plume advancement during the study period and limited Fickian spreading of mass. Observed spreading rates were generally consistent with theoretical vapor-phase dispersion. The plume mass diminished more rapidly than would be expected from radioactive decay alone, indicating net efflux from the plume. Estimates of upward 3H efflux via diffusive-vapor movement were >10× greater than by dispersive-vapor or total-liquid movement. Total vertical fluxes were >20× greater than lateral diffusive-vapor fluxes, highlighting the importance of upward migration toward the land surface. Mass-balance calculations showed that radioactive decay and upward diffusive-vapor fluxes contributed the majority of plume loss. Results indicate that plume losses substantially exceeded any continuing 3H contribution to the plume from the LLRW facility during 2001 to 2011 and suggest that the widespread 3H distribution resulted from transport before 2001.