Page 1 of 1

Band maths before L2gen

Posted: Thu Nov 09, 2023 12:11 pm America/New_York
by eli_tes
Hi, I'm using MODIS-Aqua data in Seadas (in Ubuntu 22.04) and would like to change some bands before atmospheric correction. Specifically, I want to decrease/increase reflectance in the SWIR bands and then use the SWIR atmospheric correction.

I did "Math band" but I think the results are not actually stored anywhere until I save the whole product as .dim, even when I uncheck "virtual". As soon as I store the product as .dim I can't use l2gen anymore ("WARNING!!! File /media/sf_seadasClientServerShared/AQUA_MODIS.20220520T113000.L1B.dim is not a valid input mission: Mission=null")

Is there any way to work around this?

Thanks for your help!

Re: Band maths before L2gen

Posted: Fri Nov 10, 2023 1:16 pm America/New_York
by OB.DAAC - SeanBailey
If you are looking to change the reflectance for the SWIR bands (or any band for that matter) for processing with l2gen, there is no need to modify the input L1B file. The way to do this with l2gen is to modify the gain parameter. The SWIR gains for MODIS are set to unity as the signal to noise is too low to get a reasonable vicarious gain (we can get one, but the uncertainty on it is so large as to not be different that not applying a gain).

There are 16 gain values (one for each band in the VISNIR/SWIR). The current values for Aqua-MODIS are:
gain=[0.97581,0.98510,1.00373,0.98908,0.99648,0.99128,0.98844,0.99868,1.00551,1.00462,0.99858,0.98935,1.0,1.0,1.0,1.0]

Modify the last three values to match the amount of change you desire. For example, if you desire to increase the 1240nm reflectance by 5%, change the 14th element in the above from 1.0 to 1.05. (conversely, to decease it by 5%, set the value to 0.95).

One caveat here, changing the NIR/SWIR gains will result in the need to change the visible gains as well, as the system is rather connected (changing the bands used in the atmospheric correction affects the determination of the aerosol contribution.) The current visible gains are derived using a the set of NIR gain values (748/869nm). Using the SWIR channels in the atmospheric correction, for completeness, would require rederiving the visible gains.

Regards,
Sean

Re: Band maths before L2gen

Posted: Mon Nov 13, 2023 10:03 am America/New_York
by eli_tes
Hi Sean, thank you, this is very helpful!

Could you elaborate on how to rederive the other gains? Is there somewhere I can read up on this?