Skip to content

Spatial Downscaling in R

7 messages · Michael Sumner, Bede-Fazekas Ákos, Tomislav Hengl +2 more

Original
Miluji Sb
# 
Dear all,

Please forgive my inexperience with spatial downscaling. I am interested in
spatial downscaling of global temperature to grid cell. Is there a package
in R that can perform this function?

Any help/guidance will be highly appreciated.

Sincerely,

Milu
Michael Sumner
# 
Try ClimDown package, otherwise more generally raster function disaggregate.

Cheers
On Thu, 15 Mar 2018, 06:53 Miluji Sb, <milujisb at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Miluji Sb
    

    

      
      #
    

  


  

      
Thanks again! Looking into this right now.

Sincerely,

Milu

        
On Wed, Mar 14, 2018 at 9:54 PM, Michael Sumner <mdsumner at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Bede-Fazekas Ákos
    

    

      
      #
    

  


  

      
Dear Milu,

You can use any spatial interpolation method as statistical downscaling 
approach. See package gstat for IDW (inverse distance weighted) and 
several types of kriging. In case of temperature, you might use 
elevation data from a DEM or radiation data as auxiliary variables in 
those interpolation methods (e.g. regression kriging) that can handle 
auxiliary data.

HTH,
?kos Bede-Fazekas
Hungarian Academy of Sciences


2018.03.14. 22:15 keltez?ssel, Miluji Sb ?rta:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Tomislav Hengl
    

    

      
      #
    

  


  

      
For downscaling, and especially if you work with relatively smooth 
surfaces and large rasters, I recommend using gdalwarp cubicspline 
technique e.g. (http://www.gdal.org/gdalwarp.html):

gdalwarp input.tif output.tif -r "cubicspline" -tr 250 250 -co 
"COMPRESS=DEFLATE"

If you prefer to do it via R command line then either use 
system('gdalwarp ...') or consider using 
https://cran.r-project.org/web/packages/gdalUtils/

SAGA GIS also has few similar options for downscaling 
(http://www.saga-gis.org/saga_tool_doc/2.3.0/grid_tools_0.html) and is 
efficient with large rasters (it can be easily parallelized).

T. Hengl

        
On 03/15/2018 09:34 AM, Bede-Fazekas ?kos wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Miluji Sb
    

    

      
      #
    

  


  

      
Thank you for your replies. I will try with the packages you mentioned -
they look very useful, I guess my issue is that all I have a vector of
global mean. Thanks again!

Sincerely,

Milu


On Thu, Mar 15, 2018 at 9:34 AM, Bede-Fazekas ?kos <bfalevlist at gmail.com>
wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      karsten
    

    

      
      #
    

  


  

      
Hi Milu,

if you meant you are lacking raster data sources for temperatures or
rainfall here are some I used extensively for agricultural research projects
in Africa

Rainfall (37 years of global modelled data): CHIRPS at about 5km resolution
http://chg.geog.ucsb.edu/data/index.html
ftp://ftp.chg.ucsb.edu/pub/org/chg/products/CHIRPS-2.0/

Temperature (not Global but has Africa e.g.) from FEWS
https://earlywarning.usgs.gov/fews -> data download (data since 2002
and more 

also look at http://www.ccafs-climate.org/
and http://www.ccafs-climate.org/data_spatial_downscaling/ 

Cheers
Karsten


-----Original Message-----
From: R-sig-Geo [mailto:r-sig-geo-bounces at r-project.org] On Behalf Of Miluji
Sb
Sent: Donnerstag, 15. M?rz 2018 12:55
To: Bede-Fazekas ?kos; tom.hengl at gmail.com
Cc: R-sig-geo mailing list
Subject: Re: [R-sig-Geo] Spatial Downscaling in R

 Thank you for your replies. I will try with the packages you mentioned -
they look very useful, I guess my issue is that all I have a vector of
global mean. Thanks again!

Sincerely,

Milu


On Thu, Mar 15, 2018 at 9:34 AM, Bede-Fazekas ?kos <bfalevlist at gmail.com>
wrote:

            

      
      
      
    

        
can handle auxiliary data.

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

      
      
    

        
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo