Rrs or Rhow or Rhown or Rrsn to calculate Depth invariance Index using Sen2Coral?

Which reflectance(s) do I need as an input (Rrs/Rhow/Rrsn/Rhown etc.) for Sen2Coral’s Depth Invariance Index for classification?

Hi @hisyamahm,

The combination of band 2 and 3 should work in order to compute the Depth Invariance Indexes as reported here: Depth Invariant Indices
But any combination from band 1 to band 5 should work (as stated in the documentation of the algorithm that you can find in the SNAP help) with the bands 1 to 3 are in the most useful wavelength range.
I hope this will help.


Sorry, I think I wrote the question wrong, let me clarify my situation.

I am using C2RCC for atmospheric correction, now which output reflectances should I use from C2RCC (Rrs or Rhow or normalised Rhow, or Normalised Rrs?) to be processed with Sen2Coral’s Depth Invariance Index processor?

Rrs = Remote Sensing Reflectance
Rhow = Water leaving reflectance
***n = normalised reflectance
Thank you in advance

Ok sorry I did not understand your question…
I suppose that the normailsed reflectances should work, however I am not sure if it makes physically sense or not to use any of these bands as input of the Sen2Coral Depth Invariants index algorithm.
Maybe @marpet or someone else can give me some hints?

I can also not give a definite answer. But I think the rhow should be good.
Rhow is ESA compatible and rrs is more the NASA domain. The difference is just PI. rrs = rhow * PI.
And sen2coral is an ESA tool so I think the rhow is appropriate. This should be compatible with S2 L2A data.
So this is actually the question. Does sen2Coral expect L1C or L2A data as input? TOA or BOA?
This is not well explained in the help. At least I haven’t found this information.
I also don’t know if the output of C2RCC is compatible, in terms of naming of the bands, with sen2coral.
But probably you have tried this already @hisyamahm.

In the end it might not be very important. If I understood correct the index is based on ratio of the input.

I hope my comment is not too confusing.

Thank you very much @MartinoF @marpet

Ah yes I just realised that for DII calculations it do not matter whether it’s Rrs or Rhow as it is actually just ratio of two bands… Thank you again