Dear Seniors of Geostatics I have got two simple questions that I want to include in my new manuscript. I wonder yours answers. First of all. Wiener filter is a simple kriging filter so why ordinary kriging is not a filter. Is not true that only by nugget effect the kriging estimate can exceed lower or upper limit in sample? What's the interpolation technique that is bounded by lower and upper limit in sample. Second. In the terms of correlation function on the right hand side of kriging variance we have got variance of the field. If on the left hand side of kriging variance we have got random values then we should replace random variable by random value in the expression on variance on the right hand side of kriging variance. Now, the right hand side of kriging variance would depend on unknown true (random) value and would be completely worthless. Mathematics can not lie and does not lie. Practice miners see in practice that kriging variance is not a mean squared error. If kriging variance is equal to zero value it does not mean that the estimate matches observation but only that the variance of the estimation statistics is equal to the variance of the field. It means that outcoming of input value is unknown for the mathematical model. I just wonder where I am wrong.
Best Regards t. suslo