Application of OBPG crosscalibration coefficients for MODIS Terra
Introduction
The ocean color products from MODIS Terra with the standard calibration show significant anomalies, see Franz et al., 2007. This lead the OBPG to develop a vicarious
calibration scheme, where the absolute calibration (m1 coefficients from MCST), response-versus-scan (RVS) and polarization correction coefficients are modified.
The concept and results of the OBPG vicarious calibration of MODIS Terra to SeaWiFS
have been presented in a 2008 Applied Optics paper.
Technical details regarding the retrieved correction coefficients are described in this presentation. OBPG suggests
strongly to apply both RVS corrections and polarization corrections; this requires that the user calibrates his data with the same L1B LUTs (provided below)
as were used in the vicarious calibration. It may be sufficient to use only the polarization correction, but this has not been tested by OBPG.
Please note that the results are preliminary in so far as 1) we continue to try to improve them, and 2) we do not currently update the correction on a monthly basis.
However, they do represent our current best efforts, and the application of these corrections lead to a significant improvement of the ocean color products.
Requirements overview:
1. L1A data files (required for geolocation file)
2. L1B calibration LUT (preferred) or L1B data files
3. OBPG vicarious calibration table
4. SeaDAS software (to calculate the scene polarization over oceans)
5. user provides code to apply the corrections to Terra TOA radiances
The OBPG vicarious calibration table containing the correction parameters can be obtained here
or here [currently not accessible].
The contents of this HDF file are described
in the above presentation (slide 8).
Step-by-step instructions for calculation of Q and U with SeaDAS
This section explains how to create Q and U parameters (Stokes parameters describing the polarization
state of the TOA radiance) over oceans necessary for the polarization correction. The Q and U parameters over land cannot
be calculated with the methodology used in SeaDAS.
- Download and install SeaDAS from here on a UNIX/LINUX system.
- Create a new work directory (all commands below must be issued from that folder).
- Download MODIS Terra L1a file, e.g. from here
(example file: T2007001214000.L1A_LAC; on the website, click on MODIS (Terra), then on 'Reconfigure page' to view only Terra files), uncompress with 'bunzip2 T2007001214000.L1A_LAC.bz2'.
- Either download MOD_L1Bluts.tar into work directory, extract with 'tar xvf MOD_L1Bluts.tar', or download individual 3 LUTS
(Reflective,Emissive,QA
[currently not accessible]) into work directory.
- Two options for the creation of the L1B file: 1) create it with
'modis_L1A_to_L1B.csh T2007001214000.L1A_LAC T2007001214000.GEO -rlut MOD02_Reflective_LUTs.V5.0.38.1c.hdf -elut
MOD02_Emissive_LUTs.V5.0.38.1c.hdf -qlut MOD02_QA_LUTs.V5.0.38.1c.hdf -delete-hkm -delete-qkm' (all in one command line; this
is the preferred option); or 2) download MODIS Terra L1b file from a website
(and rename it to 'T2007001214000.L1B_LAC'; this option may not reproduce the same RVS as was used in the vicarious calibration process and is therefore
not recommended).
- Save the parameter file called msl12.par into the work directory. It contains:
ifile=T2007001214000.L1B_LAC
geofile=T2007001214000.GEO
l2prod=alpha,L_q_412,L_q_443,L_q_488,L_q_531,L_q_551,L_q_667,L_q_678,L_u_412,L_u_443,L_u_488,L_u_531,L_u_551,L_u_667,L_u_678
ofile=T2007001214000.L2
- Create output L2 file with 'l2gen par=msl12.par'.
- The HDF file T2007001214000.L2 contains the Q and U components for bands 8-14. E.g. the Q component of band 8
is stored as an SDS named 'L_q_412'. Radiance unit in the output file is mW/(cm^2 sr um). These are SeaWiFS
radiance units, multiply by 10 to convert to MODIS radiance units W/(m^2 sr um). The rotation angle alpha is provided in degrees.
- Previously, I encouraged the user to confirm that his method works by calculating Q and U for MODIS Aqua granule A2002227000500 and compare
it against Q and U as presented in Fig. 9 of Meister et al., 2005. This can be beneficial for users who want to
either confirm that they use SeaDAS correctly, or for users who calculate Q and U components with alternative methods (to verify that their coordinate system is compatible
with the one used for the calculation of the polarization correction parameters). Recently, I have put a MODIS Terra L2 file (T2007001214000.L2_LAC) on our ftp site
ftp://samoa.gsfc.nasa.gov/pub/meister/xcal_sample/ for download.
It was created from the L1A file T20070012140.L1A_LAC which can be downloaded from the ocean color website: http://oceancolor.gsfc.nasa.gov/cgi/browse.pl?sen=am
(select Terra, click on reconfigure page, then go to
the date. The granule I used is located in the Pacific.)
This HDF file contains the following quantities related to band 8 (412nm) as SDS:
- Lt_412 : I_m in the 2005 AO paper (see references below)
- alpha : angle alpha in the 2005 AO paper
- L_q_412 : Q_t in the 2005 AO paper
- L_u_412 : U_t in the 2005 AO paper
- polcor_412 : p_c in the 2005 AO paper
- nLw_412 : water-leaving radiances
There are similar SDS for 443nm and 488nm. This file may be better suited for verification purposes. If the same conventions are followed as in the ocean color atmospheric
correction code, alpha, L_q and L_u, and polcor_412 can be compared. If the Q and U components are calculated in a different coordinate system, it is still possible to compare
the polarization correction factor polcor_412, because that should be independent of the choice of reference frame (as long as the angle alpha is chosen to compensate for the
difference). This picture shows polcor_412 for
scan line 1000 of the granule.
References:
- Bryan A. Franz, Ewa J. Kwiatkowska, Gerhard Meister and Charles R. McClain, The utility of MODIS-Terra for ocean color applications. Proc. SPIE, 6677, 2007.
- Ewa J.Kwiatkowska, Bryan A.Franz, Gerhard Meister, MODIS-Terra cross-calibration for ocean color bands, MODIS science team, Annapolis, Maryland, 2008.
- Ewa J.Kwiatkowska, Bryan A.Franz, Gerhard Meister, Charles R.McClain, Cross-calibration of ocean-color bands from Moderate Resolution Imaging Spectrometer on Terra platform,
Applied Optics, submitted, 2008.
- Gerhard Meister, Ewa J.Kwiatkowska, Bryan A.Franz, Frederick S.Patt, Gene C.Feldman, Charles R.McClain, The Moderate Resolution Imaging Spectrometer ocean color polarization correction,
Applied Optics, vol. 44, no. 26, pp. 5524-5535, 2005.
For further information, please contact:
Gerhard Meister, Futuretech Corp.
Ocean Biology Processing Group, NASA/GSFC Code 614.2
Building 28, Room W119, Greenbelt, MD 20771 USA
Tel #: (301) 286 - 0758, FAX #: (301) 286 - 0268
email: Gerhard.Meister@nasa.gov
Pacific.)
This HDF file contains the following quantities related to band 8 (412nm) as SDS: