Kriging...

[Chris Skelly]

Hello out there,

Micheal Shapiro suggested that I cross post this to both grassp
and grassu (This is my first posting as I have just subscribed
to these servers).  The following was more or less a recent
posting of mine to GIS-L in response to another GRASS user who
was interested in knowing if anyone else out there was writing
or had written a Kriging routine for GRASS.

If someone here has please let me know and cross post to GIS-L
if at all possible.

I am doing some research into global precipitation climatologies
and need an interpolator that will allow me to estimate areal 
precipitation from rain-guage data.  Being climatological I
have to do this in the presence of a trend.  Additionally,
I need some estimate of the spatial error involved in the
interpolation.  It appears that Kriging is my best bet so I
have embarked on creating a Universal Block Kriging routine
with multiple semivariogram model options for GRASS (This
ought to be good...as it is my first GRASS programming attempt!).

Anyway...I've started by tearing into the s.idw.surf routine which
interpolates sites data (points) via the inverse-distance-weighting
route.  The idea behind this is that the only GRASS related 
programming will be a call from the routine to a krig routine that
will substitue the idw weights with the krige weights, and some
option selection for type of model etc.

If anyone is doing similar work and/or has some suggestions for
me or anyone else doing this type of thing...I'd be happy
to hear from you!

cheers,
chris
-- 
|\/\/\/|                             Chris Skelly
|      |                             School of Earth Sciences 
|      |   I must NOT encourage      Macquarie University  
|  (o)(o)    others to FLY!          North Ryde, 2109, NSW, OZ 
C       _)                           Tel: 612-805-8433 
 |  ,___|                            Fax: 612-805-8428 
 |    /  chris at mqclimat.cic.mq.edu.au