3

I'm trying to import some Landsat 8 GeoTIFFs into GRASS (in QGIS), and I can't work out why I'm getting this error, or how to fix it. No amount of Googling has been able to tell me anything at all. I am using Landsat images as they come, straight from USGS - in other words I haven't modified the images I'm trying to import in any way. The full command output is below.

r.in.gdal input=D:_Eden\landsat\LC80910842015057LGN00_B2.TIF output=b2_GRASS -o

Over-riding projection check

Fixing subtle input data rounding error of south boundary (-3.94943e+006>1.26566e-007)

G_set_window(): North must be north of South

Finished with error

I'm running QGIS 2.10.1-Pisa with Grass 6 v2, and I installed using the Windows Standalone installer. I get the same error when trying to import the same files into the 'standalone' GRASS GIS 6 application.

This is the output I get from g.region -p:

projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 34:23:23.64S
south: 34:34:23.52S
west: 148:39:18E
east: 148:54:14.4E
nsres: 0:00:00.896576
ewres: 0:00:00.8964
rows: 736
cols: 1000
cells: 736000

Incidentally, it's a southern hemisphere location, as you'd probably guess from the above.

Does anyone have any ideas?

Improve this question
4
  • 2
    You should be careful with Landsat 8 in the southern hemisphere. NASA / USGS did some freaky stuff, and they are using northern UTM definitions for southern areas. You could try to reproject using command-line gdal before importing into QGIS to see if that fixes your problem. Try gdalwarp -t_srs epsg:<epsg-number-of-your-southern-utm-zone> <input-band> <output-file>
    – Mikkel Lydholm Rasmussen
    Aug 14, 2015 at 9:06
  • 1
    Rather than gdalwarp I would use gdal_translate -a_srs epsg:<epsg-number-of-your-southern-utm-zone> <input-band> <output-file> in order to avoid useless resampling of the data. With -a_srs you can simply fix the metadata if needed.
    – markusN
    Aug 14, 2015 at 16:24
  • Thanks... That's good to know, and may explain a few weird things I'd noticed when working with the images in QGIS - e.g. dams showing up about 30-60m east of where I expected them! Will give both gdalwarp and gdal_translate a try later today and report back.
    – Gareth Jones
    Aug 14, 2015 at 22:32
  • Out of curiosity, do you know if the same northern UTM definitions are used for ASTER data? Been doing some hydrological modelling with that and it looks mostly OK but would be good to know if I can improve the accuracy.
    – Gareth Jones
    Aug 14, 2015 at 22:34

1 Answer 1

Reset to default
1

In case anyone else is looking for a resolution for this problem, gdalwarp fixed it for me. I used the following command in QGIS and then I was able to import the image into GRASS.

gdalwarp -t_srs epsg:<epsg-number-of-your-southern-utm-zone> <input-band> <output-file>
Improve this answer
1
  • 1
    Rather than gdalwarp better use gdal_translate -a_srs epsg:<epsg-number-of-your-southern-utm-zone> <input-band> <output-file> in order to avoid useless resampling of your data. With -a_srs you can simply fix the metadata by assigning them.
    – markusN
    Aug 23, 2015 at 21:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.