Skip to content

interpolation to new points between geo coordinates

2 messages · Jan Näs, Jon Olav Skoien

Jan Näs
# 
Hi

I have a data set with geo coordinates and values for each coordinate.
I want to interpolate the values to new positions on a finer grid,
also geo coordinates.
I have looked at the fields package (interp.surface) and the akima
package (interp) but cant quite figure what I am doing wrong, or if
these functions suits my needs.

I have the two data set:

grid_1:

  lat  lon  value
1 56.5  11.1  53
2 56.6 11.1 53.1
3 56.7 11.12 52.1
4 56.5 11.2 52.9
...etc.

and a new grid

grid_2
  lat  lon
1 55.52 11.11
2 55.53 11.115
3 55.54 11.12
...etc.


And I want interpolated values for grid_2.
Any ideas?
/Jan
Jon Olav Skoien
# 
Jan,

There are a lot of packages that can help you, the best one depends on 
your needs (with or without prediction uncertainty, format of results, 
different options) and the size of your problem.
CRAN has a spatial Task View
http://cran.r-project.org/web/views/Spatial.html
with a short description of most packages dealing with spatial data. I 
think the functions you mentioned should be able to solve your problems, 
but I dont have experience with either of them. It is impossible to know 
what you are doing wrong as you did not post any error messages.
For increasing the resolution of your data, you can also try 
disaggregate or resample in the raster package. gstat, with  automap or 
intamap as simpler interfaces can also be used for geostatistical 
interpolation to the higher resolution grid, also giving you a 
prediction uncertainty. You should in general be careful with 
interpolation of lat-lon data, consider using spTransform to get 
projected coordinates if you use any of the geostatistical methods.

You will for spatial questions generally get quicker response from the 
r-sig-geo mailinglist.

Best wishes,
Jon
On 02-Jul-12 10:47, Jan N?s wrote: