Global water vapor value also used by Sentinel-2?

[griembauer]

I am creating L2 ARD using the force-level2 command from a mix of Landsat 8/9 and Sentinel-2 scenes (both are stored in the level1 folder). For now, I didn't bother to create a Water Vapor Database and only assign a global water vapor value in the parameter file, e.g.

# WATER VAPOR CORRECTION OPTIONS
# ------------------------------------------------------------------------
# This is the directory where the water vapor tables are located. Water
# vapor tables are not required for Sentinel-2, in this case DIR_WVPLUT
# may be NULL. For Landsat, it is recommended to use this functionality.
# As a minimum requirement, DIR_WVPLUT may be NULL and a global value
# for WATER_VAPOR needs to be specified. If a directory is given, 
# WATER_VAPOR is ignored. DIR_WVPLUT must contain water vapor tables.
# The 12 climatology tables must exist at least. They are used if the
# daily tables do not exist or if there is no valid daily value.
# Type: full directory path
DIR_WVPLUT = NULL
# This specifies a global value for atmospheric water vapor content in
# g cm-2. This parameter can be a dummy value to quickly process an image
# without needing to generate a water vapor database. Note that especially
# Landsat-8 is relatively insensitive to atmospheric water vapor (depending
# on wavelength), and external water vapor is not needed to process
# Sentinel-2. The error in using a dummy value is significant for the TM
# sensors.
# Type: Float. Valid range: [0,15]
WATER_VAPOR = 1.5

My question is: Is this WATER_VAPOR value applied to the Sentinel-2 atmospheric correction as well or is the standard Sentinel-2 watervapor channel used in this case (I suppose the latter would be preferable)? I can't really find the answer in the documentation

[ernstste]

Hi,

see here: https://force-eo.readthedocs.io/en/latest/howto/wvdb.html

Background

During atmospheric correction, the effect of water vapor absorption can only be corrected if we know the amount of water vapor in the atmosphere.

If you are using Sentinel-2 data only, you can stop reading. Sentinel-2 is equipped with a water vapor channel, and thus, water vapor amount can be estimated from the images.

Stefan

[griembauer]

Hi Stefan,
Thanks for your reply! I am aware of the tutorial to create a water vapor database, but for simplicity I first wanted to go with the global water vapor value.
My question is rather whether using a global WATER_VAPOR value in the parameter file overwrites the procedure of getting the Sentinel-2 water vapor from its own channel when using both Sentinel-2 and Landsat-8/9 as input during Level2 processing.

From https://force-eo.readthedocs.io/en/latest/howto/wvdb.html:
If you are using Sentinel-2 data only, you can stop reading. Sentinel-2 is equipped with a water vapor channel, and thus, water vapor amount can be estimated from the images.
--> what if it is not only Sentinel-2?

Best,
Guido

[ernstste]

Hi Guido,

okay I see where you're coming from now.
As far as I recall S2 procesing will always use the WVP value estimated from the image. Don't take my word for it though, let's have @davidfrantz confirm.

Stefan

[davidfrantz]

Hi @griembauer,

yes, the global water vapor value won't be used for Sentinel-2.

But also for Landsat, I would suggest to download the existing water vapor database, which provides a better approximation than a single dummy value - even if not up-to-date.
https://zenodo.org/records/4468701#.YBA3LuhKgmI

Cheers,
David

[griembauer]

Great! Thanks a lot for the clarification!