Extracting study site from .bil raster files, and starting the process from there. Notes and questions

[Natalia Quinteros]

Hey everyone!

I just wanted to share the method I’m using to extract information to further run it with MaxEnt. And also ask a couple of questions. 

I gathered the .bil raster files from Worldclim.org (alt, tmax, tmin, bio, etc) and from FAO the hwsd. 

I’m doing all the processing in my laptop therefore I don’t have so much disk capability. 

I tried to convert the whole .bil raster files into grids to further turn them into ascii, but the files got so large that my computer couldn’t handle the process (not enough disk space). What I decided to do is, because my study site is relatively small (parts of the US and Mexico), to extract by mask the area I needed for each BIL raster file. I’m using ArcMap 10.1

So far everything looks good. Since extracting involves storing the data in a geodatabase, ArcMap transforms the .bil raster files into raster datasets. 

I just finished working with the hwsd data following the instructions left by Michelle Greve (10/21/09) here.  With the exception of turning the whole .bil file into a grid and then add the attributes from the table, I extracted my study area first and then stored it as a raster dataset, then join the attribute table and follow the instructions until I got my soil map. 

But, I still have some doubts about this process because everything that I’ve been reading suggests the conversion of the .bil files to a grid file following the conversion into ascii, because some incompatibilities of the BIL file with ArcGIS. I’ve notice that my extractions are different from the original files (darker or lighter). 

I wonder if this is because I still need to convert them into ascii, or I have to add the ascii table to my raster files. 

I would like to know if somebody had the same experience and how did you manage to get an accurate image? 

Thanks, 

Natalia Quinteros

[Kris]

Hi Natalia,

If you think that your extractions aren't correct you could use DIVA-GIS to convert your .bil to .grd, and then .grd to ascii.  (.bil to .grd is Data>import to gridfile), then Data>export gridfile and you select arc ascii.  Then you can load the ascii into arcmap and do a ascii to raster and compare that to your other extractions.  I have had propblems importing .grd into arcmap 10.1 

Also, dont forget that you when you import .asc into arcmap you need to keep it as integer.  Perhaps this is why you get a darker version.  I would also suggest getting an external harddrive to dump all this data in (if you aren't using the geodatabase).  You can also 'download by tile' available on Worldclim, these are much smaller files for a specific region.

What files are you trying to make?  I am not familiar with HWSD, what format are those in.  Can't you cut your climate grid and use that to cut your soil grid?

So once you have made the raster (from ascii), you can do your extract by mask.  then you can converty back to ascii, raster to ascii. 

Hope that helps.

Kristin

[Kris]

On Thursday, May 9, 2013 6:08:32 PM UTC-5, Natalia Quinteros wrote:

[Natalia Quinteros]

Hi Kristin, 

Thanks for your answer. I actually tried to use DIVA-GIS to convert the .bil into .grd and then into ascii. They became so large than even working with them one by one and then storing them into an external hard drive the computer start working super slow. 

There was a researcher from Belgium who worked with the soil data, I have the feeling that the ascii file has to be joined somehow to the grid file, like the attribute table to the soil raster, don't you think? 

What I'm thinking is that maybe I have to convert this raster datasets (the extractions) into ascii, I haven't done that yet. 

Thanks for your posting . 

N.