Incremental Sampling Methodology for improved characterization of Agent Orange dioxin in Vietnam soil and sediment

The Danang airport in Vietnam was used heavily by US forces in the 1960s and 1970s. In 2018, the remediation of dioxin contamination at the airport resulting from Agent Orange use and management was completed by the US government. Generation of reliable, defensible, and cost-effective confirmation sampling data over large areas and for large volumes was a significant challenge. Traditional discrete and composite sampling methods were utilized to assess dioxin concentrations and the extent of contamination present at the airport prior to remediation. Confirmation sampling was performed after excavation of contaminated soil and sediment across the entire 28-ha site to demonstrate that remaining concentrations met the Vietnamese standards for the airport site. The volume of excavated soil and sediment was 162,567 m(3). Incremental Sampling Methodology (ISM) was used for confirmation sampling to provide a defensible estimate of the mean concentration of dioxin remaining after excavation. The use of ISM revealed that traditional methods underestimated the volume of material requiring treatment. ISM also decreased sampling variability dramatically and provided more reliable estimates of true mean concentrations in an area when compared with traditional methods. The use of ISM: 1) better captured distributional heterogeneity and decreased variability between samples from the same DU by 64%; 2) resulted in low variability between duplicate analyses of the same sample (12%), indicating a reduction in compositional heterogeneity; 3) did not underestimate contaminant levels; and, 4) increased the frequency that excavation boundaries met project goals by 61%, when compared with traditional sampling.

Automated summary

During the remediation of dioxin contamination at Danang airport, Incremental Sampling Methodology (ISM) was used to provide a more reliable estimate, better capturing distributional heterogeneity and reducing sampling variability compared to traditional methods.