kriging: export to grid

Hi Paulo,

you suggest to add conversion functions to geoR later on, but I (as the
average lazy user) would ask the following: if in the following geoR call

kc.s <- krige.conv(s100, loc=gr.s, krige=krige.control(obj=ml.s))


gr.s is of class SpatialPixels[DataFrame] or SpatialGrid[DataFrame],
could, as output, the object kc.s be of that same class? You could argue
this is incompatible, but a call like that was never supposed to work so
you could call it progress as well. And it saves me the conversion. Of
course, gr.s and kc.s could be of class SpatialPoints[DataFrame] as well.

Bests regards,
--
Edzer

Paulo Justiniano Ribeiro Jr wrote:

I've just added to the geoR tutorial page an script with examples
on converting geoR's krige.conv() outputs to
SpatialGridDataFrame and SpatialPixelDataFrame formats defined by the sp
package

The relevant link is:
http://leg.ufpr.br/geoR/tutorials/kc2sp.R

These should be encapsulated on function in a (hopefully) near future
to facilitate conversion of outputs from both, krige.conv() and
krige.bayes()


As usual, comments/suggestions/improvements etc are very welcome!

Best
P.J.


Paulo Justiniano Ribeiro Jr
LEG (Laboratorio de Estatistica e Geoinformacao)
Universidade Federal do Parana
Caixa Postal 19.081
CEP 81.531-990
Curitiba, PR  -  Brasil
Tel: (+55) 41 3361 3573
Fax: (+55) 41 3361 3141
e-mail: paulojus AT  ufpr  br
http://www.leg.ufpr.br/~paulojus



On Sun, 14 Dec 2008, Roger Bivand wrote:

On Sun, 14 Dec 2008, Nicolas Meurisse wrote:

Hello,

After kriging with the use of the  krige.conv function, I would like to
export my result under the grid format. In order to view it into a GIS.

It was suggested to me to use the writeGDAL function into the rgdal package.
However It looks like I have a problem of supported formats
(dataset object of SpatialGridDataFrame-class or SpatialPixelsDataFrame-class
) that is not the same as the one resulting from my krige.conv function
(kriging -class apparently)

I tried several function to achieve this conversion (getRasterData,
SpatialGridDataFrame, SpatialPixelsDataFrame, create2GDAL, points2grid...).
But without any success.

#my actual kriging script:
pred.grid = expand.grid(seq(170000, 130000, l = 50), seq(190000, 160000, l =
50))

library(sp)
pred.grid$var3 <- 1
coordinates(pred.grid) <- c("Var1", "Var2")
gridded(pred.grid) <- TRUE

# Create a SpatialPixelsDataFrame or SpatialGridDataFrame object first
# - forced here by including a non-coordnate column to the data frame

myPred = krige.conv(myGData, loc = pred.grid, krige = krige.control(obj.m =
myOLSMod))

myPred = krige.conv(myGData, loc = coordinates(pred.grid),
  krige = krige.control(obj.m = myOLSMod))

# Krige with the coordinates of the object

pred.grid$predict <- myPred$predict

# add the predictions (and standard errors if you like) to the
# object, they are in the correct order;
# then write out, choosing the output column as desired

tf <- tempfile()
writeGDAL(pred.grid["predict"], tf, drivername="GTiff") # not type="Byte"!
GDALinfo(tf)

# I suppose you are using the temporary file here to test your approach
# as you ought to put the output file in a permanent place if you will
# need it later

Hope this helps,

Roger

image(myPred, loc = pred.grid, col = gray(seq(1, 0.1, l = 30)))

#my conversion attempts
tf <- tempfile()
writeGDAL(myPred, tf, drivername="GTiff", type="Byte", options=NULL)
[Erreur dans function (classes, fdef, mtable) : unable to find an inherited
method for function "gridded", for signature "kriging"]

Thanks a lot,

Nicolas

--
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no

--
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of M?nster
Weseler Stra?e 253, 48151 M?nster, Germany. Phone: +49 251
8333081, Fax: +49 251 8339763 http://ifgi.uni-muenster.de/
http://www.springer.com/978-0-387-78170-9 e.pebesma at wwu.de