Page 1 of 1

l2gen and gpt processing over mid-pacific

Posted: Fri Mar 17, 2023 12:30 am America/New_York
by gisjie
I try to process data over the mid-pacific with geographical range across the 180E/180W line. For exapmple, north=22, south=12,east=-175, west=175. But the level2 product from l2gen gave extremly large x dimension. I guess l2gen thought that the longitude is from 175W to 175E. I tried to use 185 as east boudary instead, but seadas would not reconginze boundary values exceeds [-180,180]. How to handle this situation in l2gen and gpt processing?

Re: l2gen and gpt processing over mid-pacific

Posted: Fri Mar 17, 2023 10:43 am America/New_York
by OB SeaDAS - xuanyang02
I tired l2gen on AQUA_MODIS.20230220T011000.L1B.hdf with north=22, south=12, east=-175, west=175..

I got the L2 file with the southwest point (176.85, 12.91) and northeast point (-175.47, 23.65). The x dimension is 0 to 682 pixels. I don't understand what you meant by "extremly large x dimension".
Screen Shot 2023-03-17 at 10.36.50 AM.png
Screen Shot 2023-03-17 at 10.36.50 AM.png (451.28 KiB) Not viewed yet

Re: l2gen and gpt processing over mid-pacific

Posted: Sat Mar 18, 2023 9:27 pm America/New_York
by gisjie
Thanks! I tried l2gen the same Aqua image and got similar results from you. The issue should come from the GPT processing. Below is the graph xml file I used for mosaic, and the output array was 38500 x 1100.

<graph id="AQUA_MODIS">
<version>1.0</version>
<node id="MosaicNode">
<operator>Mosaic</operator>
<sources>
<sourceProducts>${sourceProducts}</sourceProducts>
</sources>
<parameters>
<variables>
<variable>
<name>chlor_a</name>
<expression>chlor_a</expression>
</variable>
<variable>
<name>Kd_490</name>
<expression>Kd_490</expression>
</variable>
</variables>
<crs>GEOGCS["WGS84(DD)", DATUM["WGS84", SPHEROID["WGS84", 6378137.0, 298.257223563]],PRIMEM["Greenwich", 0.0], UNIT["degree", 0.017453292519943295],AXIS["Geodetic longitude", EAST], AXIS["Geodetic latitude", NORTH]]</crs>
<resampling>Bicubic</resampling>
<westBound>175</westBound>
<northBound>22</northBound>
<eastBound>-175</eastBound>
<southBound>12</southBound>
<pixelSizeX>0.0090909</pixelSizeX>
<pixelSizeY>0.0090909</pixelSizeY>
</parameters>
</node>
</graph>

Re: l2gen and gpt processing over mid-pacific

Posted: Mon Mar 20, 2023 4:00 pm America/New_York
by OB SeaDAS - xuanyang02
Yes, gpt seems to have an issue with the geographic area that crosses the date line. The work around is to do mosaic twice, resulting in two files, one for the west and one for the east.

The 1st time for the region west of the date line --
<westBound>175</westBound>
<eastBound>179.9</eastBound>

The 2nd time for the east part --
<westBound>-179.9</westBound>
<eastBound>-175</eastBound>

Re: l2gen and gpt processing over mid-pacific

Posted: Mon Mar 20, 2023 8:34 pm America/New_York
by gisjie
Good idea. Thanks a lot!