Working with Landsat Collection 2 in SNAP 8.0

I am working with Landsat Collection 2 Level 1 & 2 images of same acquisition date. My ultimate aim is to make use of ‘Biophysical Processor Landsat8’. The Landsat Collection 2 Level 2 Surface Reflectance products (LC08_L2SP) are subject to be used after applying given scale factor i.e., 0.0000275 + -0.2. When Imported in SNAP 8.0 via supplied MTL file, and looking at Min and Max range of input raster, it implies that it has been already scaled and applying the scale factor would be superfluous! The range of values somewhat hovers around expected surface reflectance range!

Nevertheless, I applied the scale factor and found Min and Max going wild;
Band Math Scale Factor

This is not the case with LC08_L1TP_C2L1 product;


The Min Max values clearly advocate pending conversion from DNS to ToA reflectance which is ‘advised’ input in the ‘Biophysical Processor Landsat8’ as per associated help window;

The Biophysical Processor computes Level-2B Biophysical products from Sentinel-2 or LANDSAT 8 reflectances. From top-of-canopy normalized reflectance data,…

However, I am interested to add Sun and View angle bands from L1TP and use the resultant L2SP product (B3-B7+Sun/View angles) to assess the results. However, the Min and Max range of the L1TP angle bands seems to be with in range without applying scale factor that is 0.01.

I am literally tangled. Need guidance.

Thanks.

SNAP applies already the necessary scaling from DN values to geophysical values.
You don’t need to care about the conversion. That’s one of the services SNAP provides.
In the L2 product you have the surface reflectance values in the L1 the angle bands are converted to degree.
In the L1 radiances are provided by default.
But you can switch to toa reflectances if you want. How to do it is explained here:

1 Like

@marpet Really Fascinating to see such surprising feature in SNAP. I hope SNAP 9 will bring more ease of working.

Thank you for guidance.

@marpet Thanks for quick help. I was able to smoothly run the Biophysical Processor Landsat8 with Landsat C2L2 (SR) and now interested to supply Landsat 8 C2L1 as input to the processor.

For Landsat 8 Level 1, the referred solution to switch from radiance to ToA reflectance, I would like to confirm that isn’t like ‘forcing’ the IdePix to read the input as ToA reflectance despite input is still in radiance? I am unsure whether or not it takes care of intermediate conversion steps as described here;

Level-1 data are delivered as Digital Numbers (DN) in an unsigned 16-bit integer format and can be converted to Top of Atmosphere (TOA) reflectance or radiance using the radiometric scaling factors provided in each scene metadata file.

Will I have to manually convert radiance to ToA for each band via band math or there is other elegant way to do this in SNAP 8.0.

Thanks

If you have set the property -Ds3tbx.landsat.readAs=reflectance IdePix will receive the correct values. You only must start from original Landsat data. If you start from previously to BEAM- DIMAP converted products this property does not have any effect.

1 Like

@marpet May I deduce that by setting the property as Ds3tbx.landsat.readAs=reflectance the IdePix will automatically convert Radiance to Reflectance?

Yes, but it is not IdePix but the reader which is doing the conversion.

1 Like

Thank you for making things clear to me. :slightly_smiling_face:

1 Like

I work wih Landsat Collection 2 in SNAP 9.0. My aim is to use ‘Biophysical Processor Landsat8’ as well.
From your discussion I understood, that the only way for Collection 2 dataset is:

  • downloading Landsat C2L1 data, whereas C2L2 does not support view azimut/zenit
  • preprocessing LC8 C2L1 data by IdePix
  • using ‘Biophysical Processor Landsat8’

My issue/questions:

  • Did I understand the workflow correctly?
  • I added the configuration file to [USER_HOME]/.snap/etc/s3tbx.properties with “s3tbx.landsat.readAs=reflectance”, the imagery was processed, but before finishing IdePix following error returned: “A problem occurred during the target product processing. Type: OperatorException. Message: Waiting thread received a null tile” and the result is the imagery containing 90% of NoData
    Any suggestion what is the reason?
  • Can be used this workflow for Landsat-9 data too?

Thanks in advanced.

Hy @sitko,

Maybe this topic could help you

Thanks Diana, this workflow works well