This is a refined recipe of how to iteratively analyze COCCON open path measurements using H2O in ambient air for estimating the ILS of the EM27/SUN.
The re-evaluation of the recipe has been triggered by questions raised by Tomi Karppinen, FMI (Thanks to Tomi for this!).
The recipe takes into account that the temperature inside the EM27/SUN spectrometer is higher than outside.

The path length inside the spectrometer (from entrance window to detector) amounts ~ 74 cm.
The path length from the first tracker mirror to the entrance window amounts ~ 33 cm.

(0) total pressure is the same in and out, value is set according to measured value (and kept constant during the iteration described below)
(1) a reasonable starting H2O column value for the outside column is chosen, col_out.
(2) ratio of columns in and outside is calculated: col_in / col_out = (l_in / l_out) * (T_out / T_in), so we can init col_in also.
(3) the starting partial pressure of H2O is calculated from this choice (using col_out and measured values T_out, ptot, l_out)
(4) the H2O partial pressures inside and outside are chosen equal (and so remain equal during iterative LFT calls - they are both updated to the same new value), so ppart_h2o_in = ppart_h2o_out.

Now initialization is ready, we can iterate ...

(5) fitted column scaling factor is applied to both col_in and col_out during repeated LFT calls
(6) H2O partial pressures values are updated during the iteration together with col_in and col_out, also applying the column scaling factor (so H2O col_in and col_out both change but preserve ppart_h2o_in = ppart_h2o_out).

