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

Re: GEOSTATS: Effective range

Expand Messages
  • Syed.R.Syed@EXXON.sprint.com
    In response to Abedini s question on effective range: 1) As a rough measure, some people take the intercept between the variance of the data and the variogram
    Message 1 of 2 , Jan 20, 1997
      In response to Abedini's question on effective range:

      1) As a rough measure, some people take the intercept between the variance
      of the data and the variogram as the "effective range." Doesn't always
      work; the variance might not intercept the variogram. Also, some
      human judgment need to be applied; not any intercept makes sense (e.g.,
      power law behavior, presence of a drift, etc).

      2) You can eyeball the range; there is nothing wrong with this. This is
      probably better than (1).

      3) The minute you have "nested" structures, you potentially might have
      different heirachies of variability in your data. Maybe you can explain
      the physics behind this; maybe you wouldn't be able to. But it always
      pays to know why you have such behavior. E.g., when you have different
      ranges, are they because: of a laminae system, which makes up a bedset,
      which further makes up a sequence?

      4) Keep the model variograms simple. Using just one model usually suffices.
      No need to model every kink in the sample variograms. Get the major
      features right (early behavior, sill, and range) by using one model.
      Keep in mind the parameters that would most affect your kriging results
      and work with these parameters. Maybe you're using a small search
      neighborhood and modeling the smaller lags correctly is more important
      in your study. Maybe you wouldn't even need a "correct" range.

      5) The same applies to anisotropy. Is it really going to affect your
      predictions? How about using the same variogram and trying out an
      anisotropic search neighborhood? Use more samples in your predictions
      along the major axis of anisotropy.

      6) Why would something be anisotropic? Know the "why" is usually more
      important than modeling the anisotropy blindly. Knowing the "why"
      guarantees that your predictions would make sense. Not knowing the
      "why" doesn't guarantee anything. In fact your predictions could be
      garbage.It's always good to make sure that the "anisotropy" is not
      an artifact of the data itself.

      Regards, Syed
      --
      *To post a message to the list, send it to ai-geostats@....
      *As a general service to list users, please remember to post a summary
      of any useful responses to your questions.
    Your message has been successfully submitted and would be delivered to recipients shortly.