الملخص

ABSTRACT As it is known the process encounter the researchers who are working in spatial statistics, in mining fields about ores, ground water, and agriculture and engineering sciences and ecology. From these experiences one concludes that the data have not normal distribution or approximated to normal distribution which is usually,supposed in theoretical side of studying spatial process. In this research we discuss the robust estimation of variogram function which is considered as effective tool in studying and prediction of stochastic process and also when the stochastic process is far or approximated the normal distribution. We show here the transformation of fourth root by using with and without M-Estimation method that result real stationary estimation of variogram function. The robust variogram function is applied on real data and another real data containing outliers. The results are encouraging.