Sensitivity Analysis Of Parameters Used In The Vapor Intrusion Model For Contaminated Site Risk Assessment

We depend on soil to perform many functions. The soil, however, has been polluted by the presence of xenobiotic (human-made) chemicals or other alternatives in the natural soil environment. It is typically caused by industrial activity, agricultural chemicals, or improper disposal of waste. The contaminated soil will generate a series of ill consequences on health and economics. In this sense, it is of great significance to implement risk assessment for contaminated sites. Vapor intrusion pathway is one of the most important exposure pathways for VOCs or SVOCs in contaminated sites. The guidance of contaminated sites risk assessment in China has just been published, but the system of risk assessment management has not been completed. The basic research of vapor intrusion model which is structurally complicated and full of parameters is not sufficient and many users may not understand the model, but use it unknowingly in which case, they cannot identify the important parameters to reduce the impact of the uncertainty of the parameters on the model results with effective measures.For these reasons, based on analyzing rationale ranges of 24 important parameters for the model, the parameter analysis approach was used to discuss the physical meaning of volatile factor (VF) and the sensitivity parameters qualitatively in different scenarios. Then variance decomposition approach was used to conduct global sensitivity analysis to discuss the effect of all parameters in different scenarios, which are formed by 3 different contaminants (Koc<10 L·kg-1; 10 L·kg-1<Koc<103 L·kg-1; Koc>103 L·kg-1) from 2 different depths(100 cm; 1000 cm) below subsurface with 12 types soils(from silty clay loam to sand) to nearby buildings with convection or not. The important results are shown as follows.(1) As shown in the resuilts of local sensitivity analysis, Koc and θws are the most important parameters to impact the degree of participation and diffusion for contamination, respectively. DFia will impact the result of Ⅵ model in any scenario. However, the effect of other parameters is determined by the relative strength of the diffusion and convection process in the soil or the bottom of the building.(2) With the result of variance decomposition, the parameters are distinguished into three types to help the basic research and engineering realization, based on monitoring, fom, Ls,θ, θws, pb, Kv, based on statistics, Ab, Lcrack, LB; and based on further research, η, ER, dP,θacrack. At the same time a sensitivity parameter identification framework is constructed to identify the important parameters conveniently and simplify uncertainty analysis. (3) Based on parameter calibration, Dalian Chemical Group contaminated site was selected as an example to implement probability risk assessment of vapor intrusion pathway with the framework contributed in the paper and without. And Benzene and Benzoapyrene were selected as chemical of concern. The result shows that the error due to simplifying parameters by the framework is less than 0.2 orders of magnitude than normal, which indicates that the framework is of significance to engineering.