MBAC for MODIS

[bbbarnes]

Hello-
I'm trying to apply the MBAC for MODIS (https://doi.org/10.3389/feart.2019.00116). What are the L2GEN arguments needed for this?
Additionally, as noted in the original publication, SVC of the SWIR bands is "essential to provide reliable MBAC retrieval." But the gains for the SWIR bands are still set to 1.000 in default MODIS processing. Can you please share the gains used in the original publication, or versions with "improvements" developed in the interim?
Thanks!
-brian

[jared]

Hi Brian,

To my best knowledge, SVC has not been done for MODIS SWIR bands.

Best,
Minwei

[OB.DAACx - SeanBailey]

Well, we did produce a set of vicarious coefficients for use when we fiddled with MODIS and MBAC. Amir Ibrahim was going to dig up the gains I sent him to use (and he still might), but just in case...here are what I *think* are the ones we used:

Code: Select all

+------+--------------+
| band | avg(g.value) |
+------+--------------+
|    1 |  0.972113727 |
|    2 |  0.982880090 |
|    3 |  0.999383545 |
|    4 |  0.984962454 |
|    5 |  0.990460727 |
|    6 |  0.989051363 |
|    7 |  0.983227272 |
|    8 |  0.983937181 |
|    9 |  0.997049454 |
|   10 |  0.998231000 |
|   11 |  0.998539072 |
|   12 |  0.990400206 |
|   13 |  1.000000000 |
|   14 |  0.962189226 |
|   15 |  0.939385670 |
|   16 |  0.844858350 |
+------+--------------+

HUGE caveat here...SWIR for MODIS is really noisy, and MBAC isn't really an improvement for it. MBAC was really developed for use with the OCI instrument on PACE...where it shines Also, the processing of MODIS changed since these gains were derived (we modified the gaseous transmittances), so the gains may no longer be really valid.

Sean