Earthdata Forum

Good afternoon!

I have been working with MODIS chl-a derived from OC3, GIOP and GSM algorithms in coastal waters. I have downloaded the L1 from the website and after processed them to L2 in Seadas. The problem is that for the same image/date with no clouds, GIOP and GSM results in failed pixels or masked pixels for big areas, when OC3 does not.  I suppose it is because OC3 is empirical, and GIOP and GSM are semi-analytical and the processes are different, but I wonder if there is someway of solving it? 
Thank you very much for your time,
Have a nice day!

Ana.

Ana,

Yes, it is quite possible for the OCx algorithm to return a result while the semi-analytic algorithms fail.  The IOP algorithms require more spectral information (and consistency) than the empirical algorithms.  For the GIOP algorithm, there is a product-specific flag array that can be output to help diagnose why failures occur: flags_giop
The definition of these flags can be found in the product guide (see Appendix A).
In this case it appears that most of the failures are due to RRSDIFF which  is set if ?Rrs from Eq. 10 in the guide exceeds 0.33 (33% mean absolute difference between
modeled and observed Rrs(?).  There are a LOT of tunable parameters for GIOP... You might try increasing the giop_maxiter from the default value, but it may be that the spectral consistency of the region is insufficient (likely a calibration or atmospheric correction error) to allow a valid retrieval.

Sean