It is an average of the highest quality level observations in the compositing period (e.g., all quality 0, or all quality 1).

-- bryan

My thesis thanks you Bryan

Hi Bryan

Sorry to bug you again with something rather trivial!  I've been looking around for descriptions of the quality flags and found one by Kay Kirpatrick (2003) for 0 to 3 but notice on your own http://oceancolor.gsfc.nasa.gov/DOCS/modis_sst/ that the levels are 0 to 4.  Understanding that 0 is good and 4 is failure or land masked, is there any descriptions available for 1,2 and 3 or a weighted test description?

Meg

Meg,

We separated Kay's QL=3 into 3 (bad retrieval), and 4 (non-retrieval).  Otherwise, the QLs and quality algorithms were defined by Kay and they have not significantly changed since 2006.  I wrote the document you reference based on Miami's code, as I was porting it to the OBPG environment.  There's a bit more on the Miami vs OBPG transition here, that may help:

http://oceancolor.gsfc.nasa.gov/DOCS/modis_sst/sst_transition.html

Note that Peter Minnett and Bob Evans at Miami are still responsible for the MODIS SST algorithm and quality screening.  We just implement and process.  If you want to be sure, you can also look at the code (sst.c) that is in SeaDAS.

-- bryan