Sensitivity analysis of a contaminant transport model

Estimation and prediction of transport of pollutants following through porous media (soils/rock mass) reaching ground water and accumulating there, is becoming increasingly important because of its obvious hazard to human health. Various numerical models have been developed for prediction of pollutant transport in porous media based on advection - dispersion equation. The concept of sensitivity plays an important role in parameter estimation. It provides a means to identify the main contributors to uncertainty in the numerical model out put. In the present study sensitivity of a contaminant transport model MT3D with respect to input parameters has been evaluated. Four input parameter viz. Effective porosity, vertical transverse dispersivity horizontal transverse dispersivity and longitudinal dispersivity have been selected to check the sensitivity of this model. For this first a base simulation is carried out. Then subsequent simulations have been taken by varying each parameter by -10% -5%+5% and +10% keeping all other parameters unchanged. The final concentration of pollutants of each subsequent simulation are compared with that of base simulation Results of the analysis indicate that sensitivity coefficient of effective porosity is higher in all cases. It is found to be most sensitive parameter, which may produce more uncertainty in the prediction.