[GRASS-user] gdal_translate

GRASS GRASS user
1

Hi

Is it wrong to use -a_ullr option in gdal_translate to clip a small portion of the region from the big geotiff file region?

–
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

2

Hi

Is it wrong to use -a_ullr option in gdal_translate to clip a small
portion
of the region from the big geotiff file region?

The option -a_ullr will change the georeference of the resulting file, so
you could say it's "wrong" if you want to keep the original referencing.
The way to clip a portion of the original and still maintain
geo-referencing is with the -projwin option.

--
Micha

--
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

This mail was received via Mail-SeCure System.

3

Hi

Below is a step-by-step of what I have done but I’m getting an error when running v.rast.stats vector=catchments raster=rainfall layer=1 colprefix=area.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal in=/home/tgumede1/grassdata/Cape_Town/TRMMLast1day.tif out=rainfall

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=rainfall -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:30S
south: 33:45S
west: 18:15E
east: 19E
nsres: 0:15
ewres: 0:15
rows: 1
cols: 3
cells: 3
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal in=/home/tgumede1/grassdata/Cape_Town/Dem_CF.tif out=dem target=SRTMDEM

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=dem -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:40:46.499215S
south: 34:00:52.499215S
west: 18:17:55.500436E
east: 19:10:16.500436E
nsres: 0:00:03
ewres: 0:00:03
rows: 402
cols: 1047
cells: 420894
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.watershed elevation=dem accumulation=acc drainage=direction basin=catch stream=str threshold=200

SECTION 1a (of 5): Initiating Memory.
SECTION 1b (of 5): Determining Offmap Flow.
100%
SECTION 2: A * Search.
100%
SECTION 3: Accumulating Surface Flow.
100%
SECTION 4: Watershed determination.
100%
SECTION 5: Closing Maps.
100%
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.to.vect -s in=catch out=catchments feature=area
Extracting areas…
100%
100%
Building topology for vector map …
Registering primitives…
60653 primitives registered
314051 vertices registered
Building areas…
100%
19885 areas built
1 isles built
Attaching islands…
100%
Attaching centroids…
100%
Number of nodes: 40769
Number of primitives: 60653
Number of points: 0
Number of lines: 0
Number of boundaries: 40768
Number of centroids: 19885
Number of areas: 19885
Number of isles: 1
r.to.vect complete.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > v.rast.stats vector=catchments raster=rainfall layer=1 colprefix=precip

DBMI-DBF driver error:
SQL parser error: syntax error, unexpected NULL_VALUE processing ‘NULL’
in statement:
UPDATE catchments SET precip_min=-NULL WHERE cat=10163
Error in db_execute_immediate()

ERROR: Error while executing: ‘UPDATE catchments SET precip_min=-NULL WHERE
cat=10163’

Here is the output of gdalinfo TRMMLast1day.tif

Origin = (18.250000000000000,-33.500000000000000)
Pixel Size = (0.250000000000000,-0.250000000000000)

-------------------- coordinates--------------------------------------------
Corner Coordinates:
Upper Left ( 18.2500000, -33.5000000)
Lower Left ( 18.2500000, -33.7500000)
Upper Right ( 19.0000000, -33.5000000)
Lower Right ( 19.0000000, -33.7500000)
Center ( 18.6250000, -33.6250000)

Here is what I did to clip the region of interest.

gdal_translate -a_srs EPSG:4326 -projwin 18.2987501 -33.6795831 19.1712501 -34.0141665 3B42RT.2010032900.1day.tif TRMMLast1day.tif

Is there something I have done wrong in these steps or there is something wrong with my coordinates?

2010/7/6 <micha@arava.co.il>

Hi

Is it wrong to use -a_ullr option in gdal_translate to clip a small
portion
of the region from the big geotiff file region?

The option -a_ullr will change the georeference of the resulting file, so
you could say it’s “wrong” if you want to keep the original referencing.
The way to clip a portion of the original and still maintain
geo-referencing is with the -projwin option.

–
Micha

–

Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

This mail was received via Mail-SeCure System.

–
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

4

Hello Sandile:
It seems you are importing two raster with vastly different resolutions.
(I think we already came across this). See below...

Hi

Below is a step-by-step of what I have done but I'm getting an error when
running v.rast.stats vector=catchments raster=rainfall layer=1
colprefix=area.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal
in=/home/tgumede1/grassdata/Cape_Town/TRMMLast1day.tif out=rainfall

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map <rainfall> created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=rainfall -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:30S
south: 33:45S
west: 18:15E
east: 19E
nsres: 0:15
ewres: 0:15
rows: 1
cols: 3
cells: 3

Here, the rainfall data has a resolution of 0:15 = 15 minutes or 1/4
degree. THat's approximately (at the equator) about 27 km. So *one* raster
cell is 27 km X 27 km =~ 730 sq.km. Your region is covered by 3 cells, 1
row by 3 columns. Not very helpful data!

Next...

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal
in=/home/tgumede1/grassdata/Cape_Town/Dem_CF.tif out=dem target=SRTMDEM

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map <dem> created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=dem -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:40:46.499215S
south: 34:00:52.499215S
west: 18:17:55.500436E
east: 19:10:16.500436E
nsres: 0:00:03
ewres: 0:00:03
rows: 402
cols: 1047
cells: 420894

Your DEM layer, on the other hand, is of resolution 3 arc seconds, or
about 90 meters on a side. So each cell is 90 m. X 90 m = 8100 sq.m. =~
0.0081 sq.km.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.watershed elevation=dem
accumulation=acc drainage=direction basin=catch stream=str threshold=200

Here you chose a threshold of 200. That's 200 cells, so about 200 X 0.0081
sq.km, or 1.6 sq km. As a result (see below, the output of r.to.vect) you
are getting over 19,000 tiny little catchments. Are you sure that's what
you want??

Finally, you're trying to get raster values for 19,000 tiny vector areas
where the raster (rainfall) is only 3 cells! You'll have 1000's of
catchments with all the same values. And I guess that some of these
catchments are extending outside of the three rainfall cells, and causing
the NULL value error.

In short: I think you'll need to match the resolution of the DEM to that
of the rainfall data. If the rainfall is only as accurate as 1 data value
per 730 sq.km. then you will be able to do vector-raster analyses only at
that resolution = i.e. continent scale maps.

HTH
--
Micha

SECTION 1a (of 5): Initiating Memory.
SECTION 1b (of 5): Determining Offmap Flow.
100%
SECTION 2: A * Search.
100%
SECTION 3: Accumulating Surface Flow.
100%
SECTION 4: Watershed determination.
100%
SECTION 5: Closing Maps.
100%
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.to.vect -s in=catch out=catchments
feature=area
Extracting areas...
100%
100%
Building topology for vector map <catchments>...
Registering primitives...
60653 primitives registered
314051 vertices registered
Building areas...
100%
19885 areas built
1 isles built
Attaching islands...
100%
Attaching centroids...
100%
Number of nodes: 40769
Number of primitives: 60653
Number of points: 0
Number of lines: 0
Number of boundaries: 40768
Number of centroids: 19885
Number of areas: 19885
Number of isles: 1
r.to.vect complete.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > v.rast.stats vector=catchments
raster=rainfall layer=1 colprefix=precip

DBMI-DBF driver error:
SQL parser error: syntax error, unexpected NULL_VALUE processing 'NULL'
in statement:
UPDATE catchments SET precip_min=-NULL WHERE cat=10163
Error in db_execute_immediate()

ERROR: Error while executing: 'UPDATE catchments SET precip_min=-NULL
WHERE
       cat=10163'

Here is the output of gdalinfo TRMMLast1day.tif

Origin = (18.250000000000000,-33.500000000000000)
Pixel Size = (0.250000000000000,-0.250000000000000)

--------------------
coordinates--------------------------------------------
Corner Coordinates:
Upper Left ( 18.2500000, -33.5000000)
Lower Left ( 18.2500000, -33.7500000)
Upper Right ( 19.0000000, -33.5000000)
Lower Right ( 19.0000000, -33.7500000)
Center ( 18.6250000, -33.6250000)

Here is what I did to clip the region of interest.

gdal_translate -a_srs EPSG:4326 -projwin 18.2987501 -33.6795831 19.1712501
-34.0141665 3B42RT.2010032900.1day.tif TRMMLast1day.tif

Is there something I have done wrong in these steps or there is something
wrong with my coordinates?

2010/7/6 <micha@arava.co.il>

> Hi
>
> Is it wrong to use -a_ullr option in gdal_translate to clip a small
> portion
> of the region from the big geotiff file region?
>

The option -a_ullr will change the georeference of the resulting file,
so
you could say it's "wrong" if you want to keep the original referencing.
The way to clip a portion of the original and still maintain
geo-referencing is with the -projwin option.

--
Micha

> --
> Kind Regards
> TS Gumede
> CSIR, Meraka Institute
> 072 258 1650
>
> This mail was received via Mail-SeCure System.
>
>
>

--
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

This mail was received via Mail-SeCure System.

5

Hi

Which module do I use to change the resolutions?

2010/7/6 <micha@arava.co.il>

Hello Sandile:
It seems you are importing two raster with vastly different resolutions.
(I think we already came across this). See below…

Hi

Below is a step-by-step of what I have done but I’m getting an error when
running v.rast.stats vector=catchments raster=rainfall layer=1
colprefix=area.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal
in=/home/tgumede1/grassdata/Cape_Town/TRMMLast1day.tif out=rainfall

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=rainfall -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:30S
south: 33:45S
west: 18:15E
east: 19E
nsres: 0:15
ewres: 0:15
rows: 1
cols: 3
cells: 3

Here, the rainfall data has a resolution of 0:15 = 15 minutes or 1/4
degree. THat’s approximately (at the equator) about 27 km. So one raster
cell is 27 km X 27 km =~ 730 sq.km. Your region is covered by 3 cells, 1
row by 3 columns. Not very helpful data!

Next…

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.in.gdal
in=/home/tgumede1/grassdata/Cape_Town/Dem_CF.tif out=dem target=SRTMDEM

Projection of input dataset and current location appear to match
100%
r.in.gdal complete. Raster map created.
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > g.region rast=dem -p
projection: 3 (Latitude-Longitude)
zone: 0
datum: wgs84
ellipsoid: wgs84
north: 33:40:46.499215S
south: 34:00:52.499215S
west: 18:17:55.500436E
east: 19:10:16.500436E
nsres: 0:00:03
ewres: 0:00:03
rows: 402
cols: 1047
cells: 420894

Your DEM layer, on the other hand, is of resolution 3 arc seconds, or
about 90 meters on a side. So each cell is 90 m. X 90 m = 8100 sq.m. =~
0.0081 sq.km.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.watershed elevation=dem
accumulation=acc drainage=direction basin=catch stream=str threshold=200

Here you chose a threshold of 200. That’s 200 cells, so about 200 X 0.0081
sq.km, or 1.6 sq km. As a result (see below, the output of r.to.vect) you
are getting over 19,000 tiny little catchments. Are you sure that’s what
you want??

Finally, you’re trying to get raster values for 19,000 tiny vector areas
where the raster (rainfall) is only 3 cells! You’ll have 1000’s of
catchments with all the same values. And I guess that some of these
catchments are extending outside of the three rainfall cells, and causing
the NULL value error.

In short: I think you’ll need to match the resolution of the DEM to that
of the rainfall data. If the rainfall is only as accurate as 1 data value
per 730 sq.km. then you will be able to do vector-raster analyses only at
that resolution = i.e. continent scale maps.

HTH

Micha

SECTION 1a (of 5): Initiating Memory.
SECTION 1b (of 5): Determining Offmap Flow.
100%
SECTION 2: A * Search.
100%
SECTION 3: Accumulating Surface Flow.
100%
SECTION 4: Watershed determination.
100%
SECTION 5: Closing Maps.
100%
GRASS 6.4.0RC5+39438 (SRTMDEM):~ > r.to.vect -s in=catch out=catchments
feature=area
Extracting areas…
100%
100%
Building topology for vector map …
Registering primitives…
60653 primitives registered
314051 vertices registered
Building areas…
100%
19885 areas built
1 isles built
Attaching islands…
100%
Attaching centroids…
100%
Number of nodes: 40769
Number of primitives: 60653
Number of points: 0
Number of lines: 0
Number of boundaries: 40768
Number of centroids: 19885
Number of areas: 19885
Number of isles: 1
r.to.vect complete.

GRASS 6.4.0RC5+39438 (SRTMDEM):~ > v.rast.stats vector=catchments
raster=rainfall layer=1 colprefix=precip

DBMI-DBF driver error:
SQL parser error: syntax error, unexpected NULL_VALUE processing ‘NULL’
in statement:
UPDATE catchments SET precip_min=-NULL WHERE cat=10163
Error in db_execute_immediate()

ERROR: Error while executing: ‘UPDATE catchments SET precip_min=-NULL
WHERE
cat=10163’

Here is the output of gdalinfo TRMMLast1day.tif

Origin = (18.250000000000000,-33.500000000000000)
Pixel Size = (0.250000000000000,-0.250000000000000)

coordinates--------------------------------------------
Corner Coordinates:
Upper Left ( 18.2500000, -33.5000000)
Lower Left ( 18.2500000, -33.7500000)
Upper Right ( 19.0000000, -33.5000000)
Lower Right ( 19.0000000, -33.7500000)
Center ( 18.6250000, -33.6250000)

Here is what I did to clip the region of interest.

gdal_translate -a_srs EPSG:4326 -projwin 18.2987501 -33.6795831 19.1712501
-34.0141665 3B42RT.2010032900.1day.tif TRMMLast1day.tif

Is there something I have done wrong in these steps or there is something
wrong with my coordinates?

2010/7/6 <micha@arava.co.il>

Hi

Is it wrong to use -a_ullr option in gdal_translate to clip a small
portion
of the region from the big geotiff file region?

The option -a_ullr will change the georeference of the resulting file,
so
you could say it’s “wrong” if you want to keep the original referencing.
The way to clip a portion of the original and still maintain
geo-referencing is with the -projwin option.

–
Micha

–
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

This mail was received via Mail-SeCure System.

–
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650

This mail was received via Mail-SeCure System.

–
Kind Regards
TS Gumede
CSIR, Meraka Institute
072 258 1650