A new approach was applied to represent the natural attenuation of hydrocarbon (HC) contaminants as observed in a field lysimeter experiment. The approach describes the microbial growth on HC contaminants and linked HC biodegradation under natural precipitation conditions. The HC contaminants contained an inert component (n-dodecane) and soluble components (toluene, ethylbenzene, xylene isomers, and naphthalene), and their attenuation processes were predicted. The respiratory quinone profile method is used to indicate temporary changes in in situ microbial biomass possessed by HC-degrading microorganisms in soils. On the basis of field measurements conducted during the initial 92 days post-HC contamination, the kinetic parameters of HC-degrading microorganisms were estimated, which were used for the prediction of microbial growth and linked HC-contaminant depletion in the contaminated soil for the subsequent 259 days. The prediction was in good agreement with the measured HC concentration and corresponding respiratory quinones from HC-degrading microorganisms. The results indicate that the proposed approach can provide a reliable prediction of HC depletion in subsoil on the basis of field measurements. Further efforts are expected to incorporate the approach into a multiphase flow and multicomponent transport model for application to actual HC-contaminated sites.

Behavior of light non-aqueous phase liquid (LNAPL) was monitored in an unsaturated subsurface soil of a field lysimeter under natural precipitation condition, and simulated with a two-dimensional numerical model to evaluate the potential as source of groundwater contamination. The microbial degradation of LNAPL in the soil was greatly enhanced after 90 days from the LNAPL contamination. The model was verified by comparison withthe monitoring data of LNAPL residues after 14 and 28 days when the microbial degradation was not pronounced. However, the verified model, in which biodegradation with the first order kinetics occurred only in the aqueous phase, underestimated the microbial degradation from 28 days to 90 days, indicating the overestimation of the contaminant flux into groundwater. By assuming that the biodegradation occurred in both water and LNAPL phase in the contaminatedsoil, the verified model gave the simulation result with reasonable agreement with the monitoring data. The results suggested that evaluation of the behavior of LNAPL as source of groundwater contamination needed the appropriate estimation of the biodegradation in the LNAPL source zone. KEYWORDS; lysimeter experiment; LNAPL; numerical simulation; source zone; biodegradation evaluation;