Uncertainty analysis of hydraulic conductivity parameter in groundwater modeling using null space Monte Carlo method

In groundwater modeling, to input the hydraulic conductivity parameter, the aquifer is usually divided into some zones and for each zone a value is considered for this parameter. However, the field measurements of this parameter are performed only in some places, and therefore its generalization to other areas is associated with uncertainty. The aim of this study was to investigate the uncertainty of the hydraulic conductivity parameter in modeling the aquifer of Qom plain. For this purpose, using the MODFLOW groundwater model, the model of this aquifer was constructed and calibrated in a steady state simulation with a value of RMSE = 2.95 m for computed groundwater levels. Then the model was run using Null Space Monte Carlo to identify areas of hydraulic conductivity that calibrate the model with the same error rate as the original model. The results showed that there are eight other different zones of hydraulic conductivity that if used, the model will still be calibrated with the same error rate. The highest standard deviation of the hydraulic conductivity parameter for these eight zones was about 28 meters per day in the central and northern areas of the aquifer and the lowest standard deviation was less than 1 meter per day that occurred in the southern and eastern areas of the aquifer. This means high uncertainty of the hydraulic conductivity parameter in the central and northern areas and less uncertainty in the southern and eastern areas. Given that no significant relationship was found between the amount of uncertainties of hydraulic conductivity and the location of piezometers, so it can be concluded that increasing the number of piezometers in the model does not help to reduce the uncertainties of the hydraulic conductivity parameter and therefore more measurements of this parameter in different parts of the aquifer is needed.