Skip to content

output gstat and maps

6 messages · Marcelo Laia, Ernesto Jardim, Hisaji ONO +2 more

Original
Marcelo Laia
# 
Hi, members

I'm Marcelo, oceanographer from Brazil
and actually work with fishery acoustical data
of brazilian shelf. I'm using R, GMT, ArcView (windows)
in Linux redhat 9.0 (Athlon Xp 1.6Ghz, 20Gb 540mb mem).
I run kriging analysis using gstat (R1.8.1) in
fishery acoustic data:

x     y     sa (backscatering acoustic parameter)

the output is ordinary kriging surface.
I need overlap these output to a map of
Brazil coastal shelf and slope but i don't
make this. I have a file (ascii) of Brazil coastal,
and shapefile. Someone can helps me with this
problem? I try in gimp, kpaint and others but i
need a better solution to learn more in R.

Thanks in advance!
Marcelo
Ernesto Jardim
# 
Hi Marcelo,

If you have everything on the same coordenate system you can just plot
the Brasilian coast with plot(..., type="l") and add kriging surface
with points or something similar.

Hope this helps

EJ
On Tue, 2004-01-13 at 15:27, Marcelo Bruno wrote:
Edzer Pebesma
# 
Mmm, yeah, the trouble is that R is not a GIS, and
GMT and ArcView are no statistics packages.

I think you can import shapefiles using packages
maptools, and perhaps Rmap (not on CRAN?).
I imaging you can then plot the values with

image(values.krige)

and add polygons to the plot with the plot
commands in maptools, using the add=TRUE
argument.

levelplot, which I like much more because it
draws a legend, can currently not do this
easily (i.e., without customizing panel functions);
has anyone looked into this?
--
Edzer
Marcelo Bruno wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    
1 day later

    

      
      

        
      

    

    

      
      

        


  

        

      Roger Bivand
    

    

      
      #
    

  


  

          

      
      
      
    

        
Cell by cell operations would work, because then you don't need to know
where you are. Any other operations do need to know, and wouldn't work
without functions being written. Maybe these are best handled in GIS - you
refer to PCRaster, which is downloadable, although closed source, for
Windows. GRASS is a cross-platform alternative, and its r.mapcalc gets you
a long way - my choice would be to write a wrapper to convert the image to
the structure needed to output to a GRASS raster, and do the work there.
Perhaps it would help to be able to create a new GRASS LOCATION from
within R running within GRASS?

Roger