Loading ...
Sorry, an error occurred while loading the content.

[ai-geostats] simple kriging question

Expand Messages
  • Eric Delmelle
    Hi everybody I have a little question about simple kriging (SK). I have put a pdf document on my website www.buffalo.edu/~delmelle/sk.pdf with the formula. In
    Message 1 of 1 , Jan 26, 2005
    • 0 Attachment
      Hi everybody

      I have a little question about simple kriging (SK). I have put a pdf
      document on my website www.buffalo.edu/~delmelle/sk.pdf with the formula.

      In that formula, the estimate at a particular location is a function of
      the observed values at neighboring locations and also a function of the
      mean m over the study area. Normally, as we move away from existing
      points, the mean m weights more in the formula. Furthermore, there is no
      constraint about the weights, like in OK where they have to sum to 1.



      Now here is the problem. Suppose that we work with probabilities (P(s)>0)
      and that we know the mean over the area. The mean is positive.

      Now, if I want to estimate the probability at a particular location and if
      the sum of the weights is greater than one, the term [1- sum_{i=1}^{n}
      w_{i} ] becomes negative.

      Is it true that I may end up having a negative probability? I guess there
      will be a counter effect from the first part of the equation and w_{i}
      Z_{i} will be very large.



      Any thoughts?

      Thank you

      Eric







      **************************************
      Eric Delmelle
      Department of Geography
      S.U.N.Y. at Buffalo
      Buffalo, NY 14201
      Phone: (716) 983-2038
      URL: www.buffalo.edu/~delmelle
      **************************************
    Your message has been successfully submitted and would be delivered to recipients shortly.