Page 1 of 1

Changes in VIIRS calibration?

Posted: Mon Jul 09, 2018 10:50 am America/New_York
by fsteinmetz
Dear SeaDAS team,

In order to apply alternate atmospheric correction algorithms (Polymer and Neural Network) to SeaWiFS, MODIS and VIIRS in the frame of OC-CCI, we are using l2gen to produce so-called "L1C" files, which are obtained by applying l2gen to L1A files with the following options, in order to produce only the TOA signal and polarization correction, and bypassing application of gains and actual atmospheric correction :
l2gen ifile="..." geofile="..." ofile="..." l2prod="rhot_nnn polcor_nnn sena senz sola solz latitude longitude" gain="1.0 1.0 ... 1.0" atmocor=0 aer_opt=-99 brdf_opt=0

For VIIRS, we have recently noticed a strange behaviour when switching from SeaDAS v7.4 to v7.5, which seems to be due to a change in the following look-up table :
rsb_dynamic_lut_file=$OCVARROOT/viirsn/cal/OPER/VIIRS_NPP_CAL_RSB_DYNAMIC_LUT_v2.0.0.2c_obpg.nc
to
rsb_dynamic_lut_file=$OCVARROOT/viirsn/cal/OPER/VIIRS_NPP_CAL_RSB_DYNAMIC_LUT_v2.0.0.2d_obpg.nc

The impact of this change is a very large offset in rho_t, which varies from band to band, and is worst at 443nm, where there is an increase of about 5% (the file is V2012004140000). Do you have an explanation for this change ?

Thanks,
François

Changes in VIIRS calibration?

Posted: Tue Jul 10, 2018 9:44 am America/New_York
by OB.DAAC - SeanBailey
François,

The change in the VIIRS calibration came with the R2018 reprocessing that happened in December 2017. 
Unless you specifically updated the calibration LUTs manually, SeaDAS v7.4 would have still been using the previous calibration.
With v7.5 we made finally updated the defaults in SeaDAS to the R2018 LUT versions.

Sean

Changes in VIIRS calibration?

Posted: Tue Jul 10, 2018 10:05 am America/New_York
by fsteinmetz
Thanks for the clarification, Sean !
Now I understand that the calibration changes of R2018 are indeed very large at some bands, and are compensated by opposite vicarious calibration changes (gain(443)=1.0043 in R2014, gain(443)=0.9579 in R2018).
Cheers,
François