Goal: Export a raster from R in a format that is recognized by ArcGIS.
The file I’m working with is a .grd type file, the native raster package format usually used in R:
PEM60RAT <- raster("PEM60RAT.grd")There are a number of different raster file types and ArcGIS can only read a few of them. Using the writeRaster function you can save a raster as a different file type.
Here is the full list of file types:
| File type | Long name | default extension | Multiband support |
|---|---|---|---|
| raster | ‘Native’ raster package format | .grd | Yes |
| ascii | ESRI Ascii | .asc | No |
| SAGA | SAGA GIS | .sdat | No |
| IDRISI | IDRISI | .rst | No |
| CDF | netCDF (requires ncdf4) | .nc | Yes |
| GTiff | GeoTiff (requires rgdal) | .tif | Yes |
| ENVI | ENVI .hdr Labelled | .envi | Yes |
| EHdr | ESRI .hdr Labelled | .bil | Yes |
| HFA | Erdas Imagine Images (.img) | .img | Yes |
To chose which file type to save your raster as, write it out in the file name:
writeRaster(PEM60RAT, filename = "PEM60RAT.tif")The raster will take up more or less space on your computer depending on which file type you use. The table below shows you how much space my PEM60RAT raster used when saved as each different file type:
| File type | File size | Opens in ArcGIS | Fxnl Att. table |
|---|---|---|---|
| .asc | 64,731 KB | Yes | No |
| .img | 44,926 KB | Yes | No |
| .tif | 4,255 KB | Yes | No |
| .bil | 44,256 KB | Sort of | No |
| .grd | 3 KB | No | No |
| .sdat | 44,256 KB | No | No |
| .rst | 44,256 KB | No | No |
| .envi | 44,256 KB | No | No |
Only .asc, .img, and .tif open properly in ArcGIS. I’ve found no practical difference between them, so I would suggest using .tif because uses the least amount of space.
.bil will open in ArcGIS but doesn’t plot properly. It also will not export if you already have an .hdr file.
Although PEM60RAT has an attribute table, none of these file formats will carry the attribute table over into ArcGIS.
ArcGIS needs auxiliary files to make an attribute table for a raster. So you need to tell R to create those aux files. The clause added to the code below tells R to make a .tfw file in addition to the .tif file:
writeRaster(PEM60RAT, filename = "PEM60RAT.tif", options=c('TFW=YES'))Here is a list of the other auxiliary files ArcGIS likes to use to handle rasters:
| File Type | Purpose |
|---|---|
tfw |
ESRI World file |
ovr |
pyramid layers |
xml |
schema look and histogram |
cpg |
TIFF interpretation. |
dbf |
raster attribute table |
Unfortunately, I couldn’t get ArcGIS to recognize the attributes table.
If doing all that in R isn’t working out for you, or if you open your file in ArcGIS and it still doesn’t have an attribute table, here is how to add an attribute table onto your raster within ArcGIS.
File > Add Data > Add Data > PEM60RAT.tif
Toolbox > Data Management > Raster > Raster Dataset > Copy
Raster
Pixel type > 32\_Bit\_Signed Integer
Toolbox > data management > raster > raster properties >
build raster attribute table > PEM60RAT\_CopyRaster.This makes a copy of your raster with signed integer pixels (rather than floating point pixels) and then builds an attribute table for that copy. This attribute table has three columns, OBJECTID, Value, and Count. This is similar to ratifying a raster in R.
We have a separate file (NS_PEM_final_attributes.csv) that contains the actual attribute table for the raster. Use the Join function to attach it to your raster:
File > Add Data > Add Data > NS\_PEM\_final\_attributes.csv >
(right click raster copy) > Joins and Relates > Join > Join attributes from a table.
1. Choose the field in this layer that the join will be based on: Value
2. Choose the table to join to this layer:
NS\_PEM\_final\_attributes.csv
3. Choose the field in the table to base the join on: ValueThis adds new columns to the raster attribute table based on the data in the NS\_PEM\_final\_attributes.csv table. Rows are matched up according to the Value column, which is present in both tables.
Next, to make the colors of the land cover types:
(right click raster copy) >Properties > Symbology > Unique
Values > Value Field > CLASSFrom here you can set a color scheme or set individual colors by double clicking on the icons in the symbol column.