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

AI-GEOSTATS: Sills of Directional Correlograms

Expand Messages
  • L. Wiles
    Summary: What is the interpretation of a directional correlogram that does not reach the sill? Details: I m using correlograms to describe the spatial
    Message 1 of 3 , Feb 26, 2001
    • 0 Attachment
      Summary:
      What is the interpretation of a directional correlogram that does not
      reach the sill?

      Details:
      I'm using correlograms to describe the spatial distribution of weed
      seed banks within agricultural fields. I sampled on a square grid and
      I'm calculating correlograms for every 45 degrees. In many cases,
      the correlogram does not reach a sill of one in just one direction. The
      sill is one in the other directions.

      It is obvious from a posting of the data that the range of spatial
      dependence in that direction is much smaller than the maximum
      calculated distance of the correlogram. Also, besides not reaching
      the sill, this directional correlogram exhibits typical behavior of the
      correlograms in my study that do reach a sill of one. The direction
      that does not reach the sill is the same in most of the fields (direction
      has biological meaning in this study).

      In a couple of cases, the seed bank appears to have two different
      spatial distributions within a field with the division parallel to direction
      of the correlogram that does not reach a sill of one. When I split my
      data along the division, the directional correlograms for each part
      reach a sill of one for all directions. However, the remaining fields
      appear to have a consistent spatial distribution within the field and
      any division would be arbitrary.

      I would appreciate help with interpretation.


      ------------------------------------
      Lori J. Wiles
      Weed Ecologist
      USDA-ARS-WMU
      AERC - Colorado State University
      Fort Collins, CO 80523
      Phone: 970-491-8520
      FAX: 970-491-8247
      email: wiles@...
      ------------------------------------


      --
      * 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
    • Syed Abdul Rahman Shibli
      ... Either you have a drift or trend (non constant mean) or you truly have a variable with an infinite capacity for dispersion (fractal model). A model is a
      Message 2 of 3 , Feb 26, 2001
      • 0 Attachment
        on 26/02/01 15:45, L. Wiles at wiles@... wrote:

        > Summary:
        > What is the interpretation of a directional correlogram that does not
        > reach the sill?

        Either you have a "drift" or trend (non constant mean) or you truly have
        a variable with an infinite capacity for dispersion (fractal
        model). A model is a model is a model. What I would worry about is
        what assumptions I should be making when performing an estimation
        (e.g. kriging). Choices include using the power model as is, universal
        kriging to filter out the trend, limiting my search neighborhood and
        using just the early lag behavior, external drift kriging, etc etc.
        Structural analysis is just a means to an end. What that "end" is
        usually involves coming up with a "good" map, backed up by a clear
        understanding of the phenomena. Note that stationarity only applies
        to the search neighborhood.

        Syed


        --
        * 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
      • L. Wiles
        ... Thank you for the responses. I would like to add some information in case it might be helpful: -My data is highly skewed with usually 50% zero counts.
        Message 3 of 3 , Feb 26, 2001
        • 0 Attachment
          My original question was:
          > What is the interpretation of a directional correlogram that does not
          > reach the sill?

          Thank you for the responses. I would like to add some information
          in case it might be helpful:

          -My data is highly skewed with usually >50% zero counts. It is not
          bi-modal or multi-modal. I did a ln+1 transformation before analysis.
          -Residuals after detrending the data shows the same pattern of not
          reaching a sill in one direction (There was never much of a trend).
          -Variograms also show the same pattern of not reaching a sill in one
          direction.
          -I have looked at postings of my data over and over and do not see
          any consistent pattern for all the distributions that do not reach the sill
          in one direction.

          Additional questions: Is there a more specific interpretation than
          "non-stationarity?" If it is zonal anisotropy, what sort of phenomena
          lead to zonal anisotropy?

          Thank you.



          > Details:
          > I'm using correlograms to describe the spatial distribution of weed
          > seed banks within agricultural fields. I sampled on a square grid and
          > I'm calculating correlograms for every 45 degrees. In many cases, the
          > correlogram does not reach a sill of one in just one direction. The
          > sill is one in the other directions.
          >
          > It is obvious from a posting of the data that the range of spatial
          > dependence in that direction is much smaller than the maximum
          > calculated distance of the correlogram. Also, besides not reaching
          > the sill, this directional correlogram exhibits typical behavior of
          > the correlograms in my study that do reach a sill of one. The
          > direction that does not reach the sill is the same in most of the
          > fields (direction has biological meaning in this study).
          >
          > In a couple of cases, the seed bank appears to have two different
          > spatial distributions within a field with the division parallel to
          > direction of the correlogram that does not reach a sill of one. When
          > I split my data along the division, the directional correlograms for
          > each part reach a sill of one for all directions. However, the
          > remaining fields appear to have a consistent spatial distribution
          > within the field and any division would be arbitrary.
          >
          > I would appreciate help with interpretation.
          >
          >
          > ------------------------------------
          > Lori J. Wiles
          > Weed Ecologist
          > USDA-ARS-WMU
          > AERC - Colorado State University
          > Fort Collins, CO 80523
          > Phone: 970-491-8520
          > FAX: 970-491-8247
          > email: wiles@...
          > ------------------------------------
          >
          >
          > --
          > * 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


          ------------------------------------
          Lori J. Wiles
          Weed Ecologist
          USDA-ARS-WMU
          AERC - Colorado State University
          Fort Collins, CO 80523
          Phone: 970-491-8520
          FAX: 970-491-8247
          email: wiles@...
          ------------------------------------


          --
          * 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.