Photometry

I was wondering how synthetic photometry works.

In the optical region, filter transmission curves are usually quite
nice and smooth and they do not include the effect of atmospheric
transmission (thinking of UBVRI Bessell 1990 transmission curves) .

For near-infrared filters, things get more complicated with the
inclusion of air transmission accounting e.g. for water vapor and the
response functions are shaped quite considerably by the atmosphere.

However, also in the optical regime the atmosphere has important
contribution (with telluric lines, oxygen molecules etc...). I always
though that such effects are corrected in the published photometry on
night-by-night basis, so that's why filter transmission curves only
are enough to generate synthetic optical magnitudes.

But in the infrared case, the atmospheric contribution must is
included in the filter transmissions published in literature. Why such
a different approach? Is published infrared photometry actually not
corrected on night-by-night basis to eliminate the atmospheric
contribution?

Thanks in advance,
Wolfi