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

AI-GEOSTATS: Search options of Kriging

Expand Messages
  • Muna Mirghani
    Dear list members, I get totally different distributions when I use a Search option in kriging compared to a No-search option. The data (90 points) I use
    Message 1 of 4 , May 31, 2001
    • 0 Attachment
      Dear list members,
      I get totally different distributions when I use a "Search" option in
      kriging compared to a "No-search" option. The data (90 points) I use are
      not evenly distributed with spacing ranging from 300m to 6000m; a case in
      which it is recommended to use quadrant or octant search type to avoid
      search in one direction. However, the resulting map of the no-search is
      smoother.
      Please kindly advice me on the criteria to decide on the suitable option.
      With many Thanks.
      Muna


      --
      * To post a message to the list, send it to ai-geostats@...
      * As a general service to the users, please remember to post a summary of any useful responses to your questions.
      * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
      * Support to the list is provided at http://www.ai-geostats.org
    • Felus A Yaron
      Hi Muna, Good question. The answer depends on the type of semi-variogram that you have (especially the range - if it has). I feel that the following quotation
      Message 2 of 4 , Jun 1, 2001
      • 0 Attachment
        Hi Muna,

        Good question. The answer depends on the type of semi-variogram
        that you have (especially the range - if it has). I feel that the following
        quotation represents the current knowledge about this topic:
        " .. the neighborhood must be large enough to contain three
        observations at a bare minimum and anything beyond 25 observation
        is considered more then adequate... More precise justification for
        selection of neighborhood size depend upon the measure of
        performance achieved and fluctuates according to the nature
        of the sampling pattern and the covariance. Lacking theoretical criteria,
        experiments ... is the best alternative."
        From: Olea, R.A., 1999. Geostatistics for engineers and earth
        scientists. Kluwer, Boston, Page 63.
        See also:
        Cressie Noel (1993) Statistics for spatial data. pp. 158-160, for some
        practical suggestion, or Chiles and Delfiner (1999) Geostatistics, Modeling
        Spatial Uncertainty. pp. 201- 209 for a complete discussion.

        Good luck,
        Yaron Felus.
        Ohio-State University.

        ----- Original Message -----
        From: Muna Mirghani <mirg0933@...-berlin.de>
        To: <ai-geostats@...>
        Sent: Thursday, May 31, 2001 9:31 PM
        Subject: AI-GEOSTATS: Search options of Kriging


        > Dear list members,
        > I get totally different distributions when I use a "Search" option in
        > kriging compared to a "No-search" option. The data (90 points) I use are
        > not evenly distributed with spacing ranging from 300m to 6000m; a case in
        > which it is recommended to use quadrant or octant search type to avoid
        > search in one direction. However, the resulting map of the no-search is
        > smoother.
        > Please kindly advice me on the criteria to decide on the suitable option.
        > With many Thanks.
        > Muna
        >
        >



        --
        * To post a message to the list, send it to ai-geostats@...
        * As a general service to the users, please remember to post a summary of any useful responses to your questions.
        * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
        * Support to the list is provided at http://www.ai-geostats.org
      • Isobel Clark
        ... What is a no search option? Minimum and maximum numbers of samples in kriging depend heavily on what sort of semi-variogram you have and how much nugget
        Message 3 of 4 , Jun 1, 2001
        • 0 Attachment
          > I get totally different distributions when I use a
          > "Search" option in kriging compared to a "No-search"
          > option.
          What is a "no search" option?

          Minimum and maximum numbers of samples in kriging
          depend heavily on what sort of semi-variogram you have
          and how much nugget effect is present. In addition,
          specifying a number rather than (say) a quadrant
          search leaves you with the possibility still of very
          uneven clustered sets of samples used in the
          estimation. Kriging deals with clusters but reasonably
          even coverage is still more efficient.

          Isobel Clark

          ____________________________________________________________
          Do You Yahoo!?
          Get your free @... address at http://mail.yahoo.co.uk
          or your free @... address at http://mail.yahoo.ie

          --
          * To post a message to the list, send it to ai-geostats@...
          * As a general service to the users, please remember to post a summary of any useful responses to your questions.
          * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
          * Support to the list is provided at http://www.ai-geostats.org
        • Colin Daly
          Muna, If the no search algorithm means that all the 90 data are being used then this result (sometimes called kriging with a unique neighbourhood) always
          Message 4 of 4 , Dec 20, 2001
          • 0 Attachment
            Muna,

            If the 'no search' algorithm means that all the 90 data are being used then
            this result (sometimes called kriging with a unique neighbourhood) always
            gives smoother results than using a local neighbourhood. The reason is that
            in local kriging you will krige a part of your field using some
            neighbourhood set but other parts of the field will use different
            neighbours.You tend to see discontinuity as you go from one sector to the
            next (as one neighbour point gets replaced by another). This effect is more
            noticeable when you use small neighbourhood sets. It is reduced for larger
            sets as the points getting swapped out are the distant points and will
            usually have small weights (dependent on the variogram used - they may not
            be small for something like a gaussian variogram for example). Fianlly this
            effect is not there at all if you use all the data points in your
            neighbourhood. So with only 90 points- you should use a unique
            neighbourhood. If the program is well written then you will only ever have
            to solve the kriging equations once and will never have to do any more
            neighbour searchs -so it should run fairly fast. This can be made to run
            faster if there is a dual kriging option in your kriging program - although
            you will not get kriging variance values in this case.

            I can't be sure what 'no search' means - you will have to look it up in the
            software manual - but my guess is that it refers to unique neighbourhood if
            you are seeing smoother results.

            Colin Daly


            ----- Original Message ----
            From: "Isobel Clark" <drisobelclark@...>
            To: "Muna Mirghani" <mirg0933@...-berlin.de>
            Cc: <ai-geostats@...>
            Sent: Friday, June 01, 2001 9:51 AM
            Subject: Re: AI-GEOSTATS: Search options of Kriging


            > > I get totally different distributions when I use a
            > > "Search" option in kriging compared to a "No-search"
            > > option.
            > What is a "no search" option?
            >
            > Minimum and maximum numbers of samples in kriging
            > depend heavily on what sort of semi-variogram you have
            > and how much nugget effect is present. In addition,
            > specifying a number rather than (say) a quadrant
            > search leaves you with the possibility still of very
            > uneven clustered sets of samples used in the
            > estimation. Kriging deals with clusters but reasonably
            > even coverage is still more efficient.
            >
            > Isobel Clark
            >
            > ____________________________________________________________
            > Do You Yahoo!?
            > Get your free @... address at http://mail.yahoo.co.uk
            > or your free @... address at http://mail.yahoo.ie
            >
            > --
            > * To post a message to the list, send it to ai-geostats@...
            > * As a general service to the users, please remember to post a summary of
            any useful responses to your questions.
            > * To unsubscribe, send an email to majordomo@... with no subject and
            "unsubscribe ai-geostats" followed by "end" on the next line in the message
            body. DO NOT SEND Subscribe/Unsubscribe requests to the list
            > * Support to the list is provided at http://www.ai-geostats.org


            This message may contain privileged and confidential information. If you
            are not the intended recipient, then please notify the sender of this error.
            Any disclosure, copying, distribution or misuse of this information is
            prohibited. Copyright Roxar limited.

            --
            * To post a message to the list, send it to ai-geostats@...
            * As a general service to the users, please remember to post a summary of any useful responses to your questions.
            * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
            * Support to the list is provided at http://www.ai-geostats.org
          Your message has been successfully submitted and would be delivered to recipients shortly.