[GRASS-user] TRMM (3B43) handling

[Markus Neteler]

On Fri, Nov 19, 2010 at 8:40 AM, Markus Metz
<markus.metz.giswork at googlemail.com> wrote:
> nikos ves wrote:
>> On Wed, 2010-11-17 at 10:08 +0100, Markus Metz wrote:
>>> Another option:
>>> get binary grids from here
>>> ftp://disc2.nascom.nasa.gov/data/TRMM/Gridded/3B43_V6/
>>> read
>>> ftp://disc2.nascom.nasa.gov/data/TRMM/Gridded/3B43_V6/3B43.ctl
>>>
>>> import into a latlon location with r.in.bin -f -b north=50 south=-50
>>> east=180 west=-180 rows=400 cols=1440, e.g.
>>> r.in.bin -f -b input=3B43.100101.6.precipitation.bin
>>> output=3B43.100101.6.precipitation north=50 south=-50 east=180
>>> west=-180 rows=400 cols=1440
>>
>> i was comparing the output rasters with a png [1] a same period.
>> eg Jul_2006. it seems that in the bin [2] file N is S and vise versa.
>> The phenomenon can be seen clearly by comparing the western coast of
>> India between [1] and [2]
>>
>> I don't think I'm doing something wrong, but can anyone cross check just
>> to be sure I didn't anything stupid on my side?
>>
>
> Confirmed, that's maybe the reason why these binary grids are well hidden.
>
> Looks like there is no way around setting GCPs and using i.rectify,
> either with these binary grids or with the hdf dumps.

how about flipping it with GDAL VRT and gdalwarp? Below I cite an older
email from the GDAL list with a similar problem  + solution.

Markus

PS: From gdal mailing list:

On Fri, Jan 16, 2009 at 9:52 PM, Frank Warmerdam wrote:
> Greg Ederer wrote:
>>
>> Hi,
>>
>> I'm running gdal_translate against a VRT file:
>>
>> <VRTDataset rasterXSize="751" rasterYSize="801">
>>  <GeoTransform>-20.05, 0.1, 0.0, 40.05, 0.0, -0.1</GeoTransform>
>>  <SRS>
>>        GEOGCS[&quot;WGS 84&quot;,
>>            DATUM[&quot;WGS_1984&quot;,
>>                    SPHEROID[&quot;WGS 84&quot;,6378137,298.2572235630016,
>>                            AUTHORITY[&quot;EPSG&quot;,&quot;7030&quot;]],
>>                    AUTHORITY[&quot;EPSG&quot;,&quot;6326&quot;]],
>>            PRIMEM[&quot;Greenwich&quot;,0],
>>            UNIT[&quot;degree&quot;,0.0174532925199433],
>>            AUTHORITY[&quot;EPSG&quot;,&quot;4326&quot;]]
>>  </SRS>
>>  <VRTRasterBand dataType="Float32" band="1" subClass="VRTRawRasterBand">
>>    <SourceFilename relativetoVRT="1">10day_precip.bin.1999121</SourceFilename>
>>    <ImageOffset>0</ImageOffset>
>>    <PixelOffset>4</PixelOffset>
>>    <LineOffset>3004</LineOffset>
>>    <ByteOrder>MSB</ByteOrder>
>>    <NoDataValue>-999.0</NoDataValue>
>>  </VRTRasterBand>
>> </VRTDataset>
>>
>> I just noticed that the resulting GeoTIFF is upside down.  Is this a
>> problem with my GeoTransform values?  If so, what values would produce the
>> image right side up?
>
> Greg,
>
> I don't see any obvious problem with the geotransform values.
>
> Perhaps the image really is upside on disk.  If so, you could flip it as
> it is read by changing your values to
>
> <ImageOffset>2403200</ImageOffset>
> <PixelOffset>4</PixelOffset>
> <LineOffset>-3004</LineOffset>
>
> Best regards,