Earthdata Forum

Hi,
I have a question regarding the original projection of level 2 ocean color data (L2_LAC_OC.x.nc files) for my further workflow. After the reprojection in SeaDAS to e.g. Geographic Lat/Lon (WGS 84) the spatial resolution becomes 1.2 km. I am working in R and therefore need to know the exact projection and resolution of the original data. In the metadata information of the file, the resolution is defined as 1 km. Where is this difference in the resolution coming from?
Is there a way to find out about the exact initial projection from the metadata of the file? With this information I would be able to reproject the files in R.
Thanks a lot!

Level 1 and 2 data are not projected, they are as measured by the instrument.  The resolution attribute defines the nominal resolution of the instrument which is only truly valid at nadir.
The size of the pixels grow along-scan, so at the edge of the swath the pixels are significantly larger than 1km.  The nominally 1km square pixel at nadir becomes a trapezoid on the order of 2x4.5km

See:
https://oceancolor.gsfc.nasa.gov/docs/bowtie/modis/

VIIRS is even stranger, as it has pixel aggregation zones along the scan intended to minimize the effect of pixel growth:
https://oceancolor.gsfc.nasa.gov/docs/bowtie/viirs/

Mapping a L2 file to a regular projection can be a non-trivial task, this is why we have the l2bin and l3mapgen programs.

Sean

Hi Sean,
thank you very much for your answer! I am able to create output files with a resolution of ~2.2 km which is the result of multiple tests and variations I undertook to find the best way to generate the product I need.
So, the SeaDAS l2bin and l3mapgen processors undertake interpolations and/or fragmentations in order to create the mapped output images with a "full" coverage until the edge of the swath?
Greetings, JC

The current l2bin code treats each L2 pixel as a point in space and bins the data into the bin with the closest central lat/lon to the pixel lat/lon.
The l3mapgen code has logic that will do an area weighting when mapping  the binned data (if the interp=area option is used).

Sean

Dear Sean, thanks for the answer and explanation! Greetings, JC