Hi - I've been noodling around with qaa and noticed a discrepancy between the implementation of qaa_v6 in SeaDAS (qaa.c) and the current documentation on the ioccg.org site (www.ioccg.org/groups/software.html).  Specifically, in Step 2 of the algorithm for the case when Rrs(670) >= 0.0015 the SeaDAS implementation adds the term 0.07*(Rrs(670)/Rrs(440))^1.1 to the pure water absorption coefficient at 670.  The corresponding term in the documentation is 0.39*(Rrs(670)/(Rrs(443)+Rrs(490)))^1.14.

I haven't experimented to check how significant this difference might be, but I wondered if this was intentional or rather it was a case of the implementation not keeping up with the documentation (or vice versa, I guess).

Also, are the values of the pure water absorption and backscatter spectra tabulated anywhere in SeaDAS (or elsewhere, for that matter)?

Thanks.

Barry,

Nice find   We had to track back a bit, but got confirmation from Zhongping that indeed the code as implemented is not correct and that the IOCCG documentation is what was intended.  Way back in the vagaries of time when the v6 code was being implemented there must have been a miscommunication.  ...and no one notice for oh, 5 years or so...until you decided to noodle around   The code will be fixed, and eventually make it's way into SeaDAS.

Thanks,
Sean

Hi Sean - Thanks for checking.  I'm porting QAA to R and just wanted to be sure I got it right.  Best, Barry

Missed this one...

> Also, are the values of the pure water absorption and backscatter spectra tabulated anywhere in SeaDAS (or elsewhere, for that matter)?

Yes.  Each mission's "msl12_sensor_info.dat" file contains aw and bbw.
How those values were computed is described on our "Spectral sources and spectral integration method" page linked off our "Technical Documents" page (in case the direct link moves...)

Sean

Thanks!!  I should have found that - lots of good stuff in that file.  I wasn't looking forward to trying to do the band-integration.   Barry

Hi Sean - sorry to bother you, but what's the difference between 'hmodisa' and 'modisa?'   I'm wondering because the values for band-averaged backscatter in the msl12_sensor_info.dat files differ.  I understand that there are more bands in the hmodisa files, but I would have thought the values for the bands the two files have in common would be the same (as they are for the band-averaged water absorption).

Thanks.  Barry

The difference is modisa was not kept up-to-date after hmodisa became the default....and now with v7.5,  it's modis/aqua

A REALLY LOOONG time ago we processed MODIS as an "ocean subset".  This did not include the non-ocean,non-1km bands.
When we added the capability to process the 250m and 500m bands, it was a "new" sensor - hmodisa.  Then, after a while, we
made that the default. It took an even longer time to get rid of the split personality, but we eventually managed it and now there
is only one (well, two if you include Terra).

Sean

Ah - I haven't updated to 7.5 yet.  Will do so now.   Thanks for the explanation.   Barry