[GRASSLIST:6451] Re: Goode in Grass

[Maciek Sieczka]

Markus

Thank you for the links. I did some trial and error. Sorry it took so long. 
Please comment.

>> >>I'd like to work on MODIS imagery of Poland, Eastern Europe. I couldn't
>> >>figure out the appropriate projection definition by myself.
>>
>> >Most MODIS maps are in Sinusoidal (on MODIS sphere).
>> >Where did you get the map?
>>
>> say ftp://ftp.glcf.umiacs.umd.edu/modis/500m/Eurasia/Goodes.EUAS.2000321/
>> via the GLCF UMIACS www
>
> Looking at
> http://xserve.flids.com/pipermail/gdal-dev/2004-March/005095.html
> "... cutting the image into pieces which I seperately reprojected using
> either a
>  sinusoidal or a mollweide projection ..."
>
> which contains the link to
> http://gis.esri.com/library/userconf/proc98/PROCEED/TO850/PAP844/P844.HTM
>
>>From this I tried:
>
>     /*
>     /*   Northern Eurasia
>     /*   REGION = "02   "
> 4,   30.0000000,  65.0000000
>     -40.0000000,  40.7366111
>     -40.0000000,  60.0000000
>     -10.0000000,  60.0000000
>     -10.0000000,  90.0000000
>     180.0000000,  90.0000000
>     180.0000000,  40.7366111
> end
>
> #####################################
> dd.r02:
>
> input
> projection GEOGRAPHIC
> units DD
> parameters
> output
> projection MOLLWEIDE
> units METERS
> spheroid SPHERE
> xshift 3335846.22854
> yshift -336410.83237
> parameters
> 30 00 00
> end
> #####################################
>
> #note that we cannot cut out on the fly:
> gdalwarp -s_srs '+proj=moll +ellps=sphere +lon_0=30E \
> +y_0=-336410.83237 +x_0=3335846.22854' \
> Goodes.EUAS.2000321.band4.tif d02_europe_EUAS.2000321.band4.moll.tif

This step seems redundant to me. Is there a reason you don't reproject to
wgs84 at once? Whether I apply the first step or not, the final wgs84
layer is the same.

> #Since it's only defined for a subregion, extract the valid part
> #(not completely true for part of greenland, I think):
> gdalwarp -s_srs '+proj=moll +ellps=sphere +lon_0=30E \
> +y_0=-336410.83237 +x_0=3335846.22854' \
> -t_srs '+init=epsg:4326' -te -40 40 180 90 \
> d02_europe_EUAS.2000321.band4.moll.tif
> d02_europe_EUAS.2000321.band4.LL.tif

1. These settings you found give more-less good result. But, according to
that ESRI document, do you understand how the x,y shifts were derived?

2. As we move east or west from the central meridian in Mollweide's 
projection, there is an growing E distortion. How to calculate an optimal 
x_0 for Mollweide's, so the E distortion is minimalized in the area of my 
interest (a part of Poland)?

3. I noticed that +y_o=-316410.83237 (20000 more than in this ESRI document) 
gives a better match in the y axis (when the VMAP0 "COASTLINE" is overlayed, 
go figure). Is it me, the ESRI document or this MODIS raster in error ? Or 
maybe the VMAP0 (20 km?!). I don't have any other global data to compare to. 
But also when some wgs84 Landsat TM of Poland is overlayed on the MODIS 
reprojected to wgs84, they fit better when "my" +y_o=-316410.83237 is used.

4. There is Goode's projection in proj 4.4.9 (+proj=goode) which takes care 
of all the "zones" in some automated way. In case of the Eurasian MODIS we 
discuss here it works, as expected, similarly to +proj=moll, but:

a. It handles the N and S utmost parts of my MODIS image properly while 
Mollweide's introduces a vast distorion there.
b. y_0 is intended not to be provided in the proj definition for +proj=goode 
(I guess), but, following my observation of the 20km vertical shift, if 
y_0=20000 is given, the resulting wgs84 MODIS fits the VMAP0 data *much* 
better than when y_0 is omitted. Example:

gdalwarp -s_srs '+proj=goode +ellps=sphere +lon_0=30E +towgs84=0,0,0,0,0,0,0 
+x_0=3335846.22854 +y_0=20000' -t_srs '+proj=latlong +ellps=wgs84 
+towgs84=0,0,0,0,0,0,0' -te 7 33 34 59 Goodes.EUAS.2000321.band3.tif 
modis_goode_esri_y0+20000.tif

Yours badly puzzled,
Maciek 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: modis&vmap0_y_no_offs.jpg
Type: image/jpeg
Size: 60025 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-user/attachments/20050413/19f2adbb/modisvmap0_y_no_offs.jpg
-------------- next part --------------
A non-text attachment was scrubbed...
Name: modis&vmap0_y_20km_offs.jpg
Type: image/jpeg
Size: 59911 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-user/attachments/20050413/19f2adbb/modisvmap0_y_20km_offs.jpg