Anomalies after gpt reproject of MERIS FRS L2

Please enter here to ask a question about any NASA Science related topics!
Post Reply
OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 221
Joined: Mon Apr 07, 2008 4:40 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB SeaDAS - knowles » Wed Jan 11, 2017 12:15 pm America/New_York

Hi Andrew,

I remember looking at this post a few months ago and I believe I wasn't able to replicate the results.  Sorry, for no follow up, but I guess it got lost in the shuffle, and each of us thought someone else was looking into it :(.  I'll take a look at this again.  Since I need to run l2gen, can you tell me whether you did anything special like extracting?  Obviously it's a big file.  Also if I can't replicate this, can you?  In other words, if you run it again, are the nan blocks occurring in precisely the same place in the image each time?

thanks
Danny

Tags:

OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 221
Joined: Mon Apr 07, 2008 4:40 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB SeaDAS - knowles » Wed Jan 11, 2017 1:11 pm America/New_York

Hi Andrew,

I ran l2gen on the whole file.  I then cropped it in the GUI to approx your dimensions.  I then reprojected in the GUI (Transverse Mercator, central meridian=-153, lat of origin=70)  For Output Parameters I set Pixel sizeX=300 and Pixel sizeY=300.  I used Nearest Neighbor.  I did not get any NaN blocks.  I'm attaching the image. 

I will run l2gen with subset params set to yield a smaller file and see if this pathway will yield the NaN blocks. 

Danny

OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 221
Joined: Mon Apr 07, 2008 4:40 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB SeaDAS - knowles » Thu Jan 12, 2017 11:02 am America/New_York

Hi Andrew,

I ran l2gen creating an extract and it yielded the NaNs problem similar to what you are seeing.  See attached image with NaNs shown in red.  In the meantime, as shown in my previous post, if you first do a spatial subset (crop) in SeaDAS of the l2 file, then the Reproject tool will work correctly.  This is related to a known issue where directly reprojecting a l2 file in far northern and far southern latitudes can cause what looks like NaN splits in the image.  We will look into getting this fixed.

thanks
Danny

andrew.meredith
Posts: 24
Joined: Tue Dec 31, 2013 10:29 am America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by andrew.meredith » Thu Jan 12, 2017 11:39 am America/New_York

Danny,

Thanks for looking into this. I apologize for omitting the critical info that i was subsetting via l2gen. Your note that it is specific to the far northern and southern latitudes which would explain why we haven't seen it previously (although would you consider WA state far north?).

Thanks for the suggested work around. We'll work with that for now but look forward to a fix down the road as it does introduce some significant inefficiencies when process a large number of files.

Andrew

OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 221
Joined: Mon Apr 07, 2008 4:40 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB SeaDAS - knowles » Thu Jan 12, 2017 12:00 pm America/New_York

Hi Andrew,

My observations of this problem have only been for latitudes a bit above 60 and below -60 degrees.  Unless WA has detached itself and gone adrift, then no I would not consider WA to be far north :)   We'll get on this one, and I expect we will get it resolved soon.

thanks
Danny

andrew.meredith
Posts: 24
Joined: Tue Dec 31, 2013 10:29 am America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by andrew.meredith » Wed Sep 19, 2018 12:01 pm America/New_York

Hi,

I just wanted to point out that this problem still exists in v7.5 and is noted with OLCI data also as seen in the attached (left side image) scene from central Florida. NaN's are masked as red. The masking within clouds is OK because the "applyValidPixelExpression" was set when reprojecting in Seadas. Reprojection in Snap (v6) with the S3tbx toolbox disabled (forcing use of the NetCDF product reader) does not produce anomalies (right side image in attached).

Is there are problem not using the seadas L2ProductReader?

I see the problem reprojecting MODIS l2gen output also.

Thanks
Andrew

OB.DAAC - SeanBailey
User Services
User Services
Posts: 1171
Joined: Wed Sep 18, 2019 6:15 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB.DAAC - SeanBailey » Sat Sep 22, 2018 2:36 pm America/New_York

Andrew,
True, we have not addressed that particular reprojection issue.  You can always override the L2ProductReader by selecting File->Import->NetCDF (Generic)

Sean

andrew.meredith
Posts: 24
Joined: Tue Dec 31, 2013 10:29 am America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by andrew.meredith » Mon Sep 24, 2018 10:30 am America/New_York

Thanks Sean.

Is it possible to override the product reader when using gpt? It looks like it may be possible in SNAP but I didn't see anything obvious in SeaDAS.

OB SeaDAS - knowles
Subject Matter Expert
Subject Matter Expert
Posts: 221
Joined: Mon Apr 07, 2008 4:40 pm America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by OB SeaDAS - knowles » Mon Sep 24, 2018 2:00 pm America/New_York

I don't think you can override the reader in gpt.  If you can open and export the level2 file as a generic NetCDF this would then work.

You could do this in a gpt workflow (see below).  The saved file then should no longer be in level2 file format but instead a generic NetCDF-CF format.  You might be able to extend the workflow to read the new file. 

For example:
gpt Workflow.xml -p WorkflowEx01.par -Pfile=A2010283180500.L2_LAC_SST.nc

Workflow.xml:
<graph id="WorkflowEx01Graph">
  <version>1.0</version>

    <node id="ReadNode">
      <operator>Read</operator>
      <sources/>
      <parameters>
        <file>${file</file>
      </parameters>
    </node>

    <node id="WriteFileNode">
      <operator>Write</operator>
      <sources>
        <source>ReadNode</source>
      </sources>
      <parameters>
        <file>${file}_generic.nc</file>
        <formatName>NetCDF-CF</formatName>
      </parameters>
    </node>
</graph>

Danny

andrew.meredith
Posts: 24
Joined: Tue Dec 31, 2013 10:29 am America/New_York

Anomalies after gpt reproject of MERIS FRS L2

by andrew.meredith » Tue Sep 25, 2018 2:58 pm America/New_York

I'll give that a try. Thank you.

Post Reply