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

GEOSTATS: co-kriging, response

Expand Messages
  • Gunther Schmidt
    Hello again, Thank you all for your comments and suggestions in your replies at my question (s.3.) dealed with co-kriging of precipation and fog data in
    Message 1 of 1 , Feb 6, 1998
    • 0 Attachment
      Hello again,

      Thank you all for your comments and suggestions in your replies at my
      question (s.3.) dealed with co-kriging of precipation and fog data in
      addition with an elevation model (the response was like kicking off an
      avalanche). In condensed form the results were as follows:
      -------------------------------------
      1. expert discussion

      Many people recommended including elevation as external drift rather
      than co-kriging with precipation values (because of the non-
      stationary behaviour of elevation models). Additionally strategies are
      shown below. Interpolating fog data seem to be in a virginal status.
      Perhaps, there will be some results soon.

      Dan Cornford <d.cornford@...> wrote:
      >However I would send some caution - why do you want to cokrige
      >elevation and rainfall? Over Germany elevation will not be
      >stationary (in mean and covariance) so estimating a variogram for
      >elevation will be difficult and estimating a cross variogram even
      >worse!!! Often by including the elevation in a regression model
      >rather than full cokriging the variogram of the residuals is rendered
      >stationary (is the rainfall variogram truly stationary - I really
      >doubt it!!! -because rainfall is strongly correlated with elevation
      >and this I am sure is non-stationary!). Another method which may be
      >more relavant to you is kriging with external drift (eg. Wackernagels
      >work with Gordon Hudson in the IJC) although inference of the
      >variogram is rather tricky under this model!

      Vijay Sathya <Vijay.Sathya@...> wrote:
      >I'am working with meteorological variables as well and as far
      >as introducing the altitude corelation.. the best way out is to use
      >altitude as external drift.

      David Garner <garner@...> wrote:
      >One idea is to also treat time as an external drift to remove
      >seasonal variations. This has been done for pollution trends over
      >cities. Then you would have 2 external drift variables: elevation
      >and time.

      Edzer J. Pebesma <e.pebesma@...> wrote:
      >Besides cokriging it can also do universal kriging (i.e. regression
      >with kriging on the residuals) which may be a good alternative if
      >your elevation data are available at every location (as a DEM or
      DTM).
      -------------------------------------------------
      2. Software (in random order); shame on me, next time I'll 1st visit:
      http://curie.ei.jrc.it/software/index.htm - the software index of ai

      Stanford University
      ftp://geostat1.stanford.edu/gslib/
      ftp://banach.stanford.edu/gslib/
      http://ekofisk.stanford.edu/SCRF.html
      GSLIB

      CRES, ANU (Australian International University), Canberra
      http://cres20.anu.edu.au/software/index.html
      ANUSPLIN

      Centre of Geovariances Fotainbleau, GEOVARIANCES
      http://www.geovariances.fr/software/isatis/index.html
      http://www.geomath.com/HomePage/software.htm
      ISATIS

      Universit├Ąt Amsterdam
      http://www.frw.uva.nl/~pebesma/gstat/
      GSTAT 2.0

      Department of Oceanography
      Texas A&M University
      http://www-ocean.tamu.edu/~baum/software-listing.html

      University of Leuven in Belgium
      www.agro.ucl.ac.be/biom-pub/www/man0.htm
      climexe.zip
      ---------------------------------------
      3. My posted question

      >I'm trying to interpolate climatic data like precipation values and
      >fog-duration measured over the whole year and for every month. The
      >exploration area is the entire territory of germany so there is a
      >large amount of input data (> 1500 values). First I used variogram
      >analysis and ordinary kriging for geostatistical examination (made
      >with FUZZEKS). In fact I've got quite good results, but I think there
      >will be much better results if I could include elevation data
      >additionally. For this case I would like to use the co-kriging
      >method. So my question is: Does anybody know or have a
      >(freeware-)software that allows me to apply co-kriging to my dataset?
      >Another question: Does anybody have further experiences in
      >interpolating fog-data?

      Thanks again,
      Gunther Schmidt





      -----------------------------------------------------------
      Dipl.-Geol. Gunther Schmidt
      Geographisches Institut
      Christian-Albrechts-Universitaet Kiel
      Ludewig-Meyn-Str. 14
      D-24098 Kiel

      Tel.: 0431/880-3432
      FAX : 0431/880-4658
      e-mail: schmidt@...-kiel.de
      --
      *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.
      *To unsubscribe, send email to majordomo@... with no subject and
      "unsubscribe ai-geostats" in the message body.
      DO NOT SEND Subscribe/Unsubscribe requests to the list!
    Your message has been successfully submitted and would be delivered to recipients shortly.