MODIS 250m Processing

[avmehta]

Hello,
    I want to get MODIS L2 data at 250 m resolution.   I am trying to first use SeaDAS GUI to process MODIS L1 to L2 assuming that I can get L2 data at 250m from L1B file in GUI! I have an L1A file and started with first getting GEO file. but I get the following error:
     OBGP session started
     Downloading 'PM1ATTNR.P2010297.1800.003' to /Users/avmehta/Applications/seadas-7.5.3/ocssw/var/anc/2019/297
      ERROR: Authentication Faulire retrieving: oceandata.sci.gsfc.nasa.gov/ob/getfile/PM1ATTNR.P2010297.1800.003
      Please check that your ~/.netrc file is setup correctly and has proper permission
    
  But my .netrc file is set correctly! I download data all the time!

  I'd appreciate your help.

  Thanks very much.
Amita

[OB.DAACx - SeanBailey]

Did you try again?   It may have been an intermittent network issue.  The script reported what it saw in the status code (Authentication Faulire) and simply suggests the most checking the most common issue that can result in that error.  If you know that the .netrc is not the issue, then that is not the issue.  You can also run modis_atteph.py from the command line to retrieve the necessary files for geolocation, or even run the full geolocation from the command line - you can then still use the GUI to process the scene to L2.

Regards,
Sean

[avmehta]

Hi Sean,
  I will try again in  GUI. In the meantime I tried to run on my linux server to get a GEO file:

modis_GEO.py A2019297182500.L1A_LAC
I get  the following error message:
Output file de200.eos EXISTS
1
Error creating binary planetary ephemeris file

In the forum I see that a few posts are there about 'de200.eos  exists' -- but I can not even find this file if it exists!

Thanks for your help.
Amita

[gnwiii]

The file should be $OCSSWROOT/share/modis/static/de200.eos.  On a 64-bit linux system:
ls -l $OCSSWROOT/share/modis/static/de200.eos
-rw-r--r--. 1 <seadas_user> <seadas_group> 4767192 Sep 24  2019 /home/<seadas_user>/ocssw/share/modis/static/de200.eos

This is a binary file that uses a system-dependent format.  The software attempts to check that the file has the appropriate format.  There are many reasons for the error you saw, including mixing seadas version, a seadas installation copied or shared from a different system, abusing root privileges (sudo, etc.), or a non-linux filesystem (USB drive or network share), or lack of space on the disk.   Note that some of the previous posts were for an older version of seadas that used a different directory structure (in particular "run" was replaced by "share".

[avmehta]

Thanks George.
I checked the directory $OCSSWROOT/share/modis/static/ but I only see de200.dat.
From the reasons you have listed, I think it may be about privileges! I am running this on a container on adapt where Seadas/OCSSW were installed for us! Perhaps I have no permission to write in $OCSSWROOT! That would be a problem right?
But the error says de200.eos files exists! That is confusing.
====
OK so the script DOES work on my server where I do have write permission!
Thanks.
Amita

[gnwiii]

Did you run "source $OCSSWROOT/OCSSW_bash.env"?   What do you get from "echo $PYTHONPATH"?  

You can easily check permissions in linux using "ls -l <filename>" and "ls -ld <directory_name>".   There can be additional "security context" if your system uses selinux (RHEL or Fedora), but that doesn't usually cause problems for user installs of the OCSSW software.

[avmehta]

I did check - I can not right/save anything on $OCSSWROOT.  Not sure how I can get around this. I assume that  de200.eos is created every time I run modis_GEO.py. Is there a way to save it in some other local directory and let the script know about it?

A related question: When I cerated MOIDS L1B file on my server, I get the following files with 250m, 500m, and 1km resolutions:

A2019297182500.L1B_HKM
A2019297182500.L1B_QKM
A2019297182500.L1B_LAC

I tried to use the QKM file to in l2gen (in GUI) -- but that did not work.
I understand that I need to use the 1 Km file and set resolution to 250 m - I am trying that next. So my question is what does the A2019297182500.L1B_QKM file have?
I appreciate your help and patience with my questions.
Thanks.
Amita

[OB.DAACx - SeanBailey]

The de200.eos file is generated only once.  It is a binary file and needs to match the architecture of the system you're running on.  While it is possible to move the ___location of the planetary ephemeris file (that's what the de200.eos file is), it would require a number of changes - not hard, but also not supported. 

L1B_LAC contains the 1km bands
L1B_HKM contains the 500m (half kilometer) bands
L1B_QKM contains the 250m (quarter kilometer) bands.

The l2gen program reads all three files, but will only accept as input the 1km file (L1B_LAC).  The output resolution is determined by the resolution parameter NOT the input file.
This is a MODIS-only option, resolution has no impact for any other sensor.  

Sean

[avmehta]

Thanks much Sean.
Perhaps I will request the system admin to run modis_GEO.py once so that de200.eos gets created! Then it  might work for subsequent processing I hope.
Amita

[avmehta]

Hello,
  I generated MODIS L1B files and GEO file on my server and generated L2 file on my Mac via GUI! I got the L2 file and I could visualize layers. However, when I closed the file and tried to open it  later  I get Java Memory error!

An exception error occurred:
Type: java.util.concurremt.ExecutionException
java.lang.OutOfMemoryError: java heap space

Earlier I had java memory issue with reprojection and I  changed it in vmoptions -- so that is already at max.
Is there anyway to get around this? The error not occur when I executed l2gen and got the L2 file, but started after I closed the file.

Thanks.
Amita