geostatistical stochastic simulation

Soil saturated hydraulic conductivity is a key parameter needed in many projects including drainage. So it necessitates knowing about the spatial distribution pattern of hydraulic conductivity. However, to obtain the knowledge, it is needed to have a lot of field measurements carried out which is time consuming, tedius, and costly. Different types of kriging can be used for estimating and mapping hydraulic conductivity over a study area. However, the estimated results contain some uncertainties. Unlike kriging, stochastic simulation can be used to model the estimation uncertainty and incorporate it into the decision-making processes. In this paper, Sequential Gaussian Simulation (SGS) and non-parametric Sequential Indicator Simulation (SIS) approaches were employed to model the uncertainty attached to the hydraulic conductivity estimates in KheirAbad plain in Khozestan. A number of 200 equally probable simulated maps of hydraulic conductivity were generated through either of the methods. The results revealed that unlike the kriged map, the simulated maps could reproduce the histogram and semivariogram of the raw data, reasonably well. Regarding local uncertainty, the results showed that the kriging variance does not depend on the actual data values and so there is a limitation in its use. The accuracy plot and width of probability interval plot indicated that the uncertainty model obtained through SGS is more accurate than that obtained through SIS; however the goodness coefficient was slightly smaller for SGS (0.88) than for SIS (0.94).