ABSTRACT
Surficial CO2 efflux surveys have been used to delineate hydrocarbon source zones in contaminated aquifers and provide estimates of hydrocarbon biodegradation rates. This approach requires distinguishing between CO2 derived from petroleum degradation and CO2 produced from natural soil respiration. To this end, radiocarbon has been used to differentiate between 14C-depleted CO2 from hydrocarbon degradation and 14C-enriched CO2 from natural soil respiration to effectively quantify the contribution of each source to total CO2 efflux, and by deduction natural source zone depletion (NSZD) rates. In this study, a systematic method comparison has been conducted to evaluate available approaches for collecting CO2 gas samples for radiocarbon analysis used to correct total CO2 efflux measurements for quantifying natural source zone depletion rates. Gas samples for radiocarbon analysis were sampled from (i) dynamic closed chambers (located at ground surface), (ii) static chambers (also at ground surface), (iii) shallow soil gas probes (0.3 m bgs), and (iv) soil gas monitoring wells (~0.6 m below ground surface) during a CO2 efflux survey conducted at the site of a historical pipeline rupture near Bemidji, MN. The mean fraction of radiocarbon (F14C) obtained from samples overlying the source zone were (i) 0.93 ± 0.01, (ii) 0.73 ± 0.03, (iii) 0.71 ± 0.04, and (iv) 0.41 ± 0.06, for the four methods respectively. These F14C values were used to apportion total CO2 efflux measurements into contributions of contaminant-derived CO2 efflux and natural soil respiration to evaluate natural source zone depletion processes. Results suggest that the method of radiocarbon sampling has a significant effect on the calculated fraction of the CO2 efflux originating from contaminant-related soil respiration, with contributions ranging between 27% and 59% of total soil respiration. Results indicate that radiocarbon sampled from static chambers and shallow soil gas probes methods offer the best compromise between CO2 sample yield and sample representativeness, providing the most reliable estimates of CO2 effluxes originating from contaminant degradation. However, the results also show that for this study, all methods agree within a factor of <2.3 regarding the inferred NSZD rates.
