Modeling Chlorobenzene Leaching from a Landfill into a Soil Environment Using Particle Filter Approach

Abstract

With increased use and public awareness of landfills, there is much concern for the pollution potential of the landfill leachate. As leachate migrates away from a landfill, it may cause serious pollution to the groundwater aquifer as well as adjacent surface waters. Modeling the behavior of pollutants during the flow of leachate through soil is essential for predicting the fate of these pollutants and designing restoration strategies. In this paper, Particle filter approach is introduced for the estimation of chlorobenzene leachating from a landfill into a soil environment. The discrete advection transport equation is used as a deterministic model, in which the state vector denotes the concentrations at discrete nodes. First, random Gaussian errors are added to the transport equation to become the state equation. Then, the Particle filter recursive process is applied for that state space model. In each time step, random samples which represent the density of state vector are propagated through state equations and updated by observation data assimilation. The result shows that the filtered estimation from Particle filter approach is in close agreement with the true value. Furthermore, the efficiency of Particle filter approach is compared with the numerical method using root mean square error (RMSE). The RMSE value for the numerical method was seven times larger than that for the Particle filter approach.

Keywords: landfill leachate; particle filter; state space model; subsurface transport model; state estimation