On Mon, 31 Aug 2009, Julian Burgos wrote:
Dear list, Hopefully this question will not be too basic. I have a relatively large dataset (~250000 rows) consisting of lat, long and a data value. The data has been collected in a regular grid, but the grid is "tilted". In other words, along a row the latitude values increase constantly, and along a column the longitude values do the same. The tilt is very slight, but large enough to cause problems when trying to convert the SpatialPointsDataFrame object holding the data to a SpatialPixels objects. My ultimate objective is to convert this data to a raster file and input it to GRASS. Any ideas on how to solve this problem? Is is possible to create raster images with a "tilted" grid?
Given the format of the input data, I don't think that GMT or GDAL will help directly. GMT xyz2grd seems to want a non-tilted grid. GDAL's gdalwarp could re-orient an input tilted raster file, but this isn't what you have. Have you considered using GRASS directly, with v.in.ascii and format=point? Then one of the v.surf.* to interpolate to a non-tilted raster - there are a few more on the GRASS Addons wiki site. That might be more direct that doing the interpolation in R, although it would come to the same thing. Hope this helps, Roger
Many thanks, Julian
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