2009
Data before day 274, 2009 (2009 Oct 1) did not behave like AOD spectra especially at shorter wavelengths.
Between day 274 and day 317 (2009 Nov 13) AOD data looked okay in comparison to NIMFR.
Then no AOD data until day 322 where the oscillations in AM are obvious until about 1030, but good after that.
About the same until day 331.
On day 331 don't trust data before 1200.
START HERE!
On day 332 don't trust data before 1200.
On day 334 data questionable before 1400.
On day 338 data data look good after 1030.
On day 362 looked like bad before 1130.

2010
On day 4 looked like bad before 1130.
On day 8 looked like bad before 1200.
On day 61 still looked like bad before 1200.
On day 88 looked like bad before 1000.
On day 120 looked like bad before 1000.
From day 117 to day213, 500-nm aod did not agree in wavelength dependence, perhaps,
due to bad determination of Vo.
Until the end of the year, things looked pretty good.

2011
Until day 42 pretty good shading and agreement.
Day 42 had shifting wavelength pattern around 1300.
Day 44 may have recovered.
Day 95 has strange wavelength dependent AOD.
Day 96; it recovered.
Some large differences on day 134 thru 281, especially pre-noon.

After response dropped, around 11 October 2011 (day 281), AODs agree quite well with direct pointing 
AOD measurements; CIMEL and NIMFR comparisons made.