See issue #437 for details

email of 2021 May 5
In the last few days, I've been collecting rates for
O i +e inelastic collisions. I started with Barklem,
A&A 462, 781-788 (2007) and tried to cast the rates into
Omega_ij strengths. After accounting for the partitioning
into the 3 fine-structure levels (lumped together by
Barklem), I was in a position to compare my Omega_ij with
those reported in the stout datafile. I was able to match
the Omega_ij for all transitions from the 1,2,3 levels
into the 4,5,6 levels, and also the 4-->5 transition.

I'm not able, however, to reproduce the Omega_ij values reported
in stout for the 1,2,3-->7 transition (from O(3P_J=2,1,0)
into the level at 9.52 eV). I didn't check the transitions
into more energetic states. The comments at the bottom of the
stout o_1.coll file (attached here for your reference) note
that there may be some inconsistencies. I'm also attaching the
values that I estimated myself for transitions within levels up
to 10 eV. I would appreciate your comments on this difference.
If you think it may be useful, I can send you a more extended
table with the Omega_ij for additional transitions.

Thank you!

   Antonio García Muñoz

Dr. Habil.
Planetary/Atmospheric Scientist
IRFU/DAP/LDE3
CEA Paris-Saclay
Bât. 703 Orme des Merisiers
91191 Gif-sur-Yvette Cedex, France
http://antoniogarciamunoz.wordpress.com/

Previous implementation had been intended to ignore values close to
underflow, but the number chosen was large enough to be a significant
source for low abundance species.  Tightening the criterion to cut the
step length if any negative abundance appears works fine after wide
testing, so seems like a better approach.

Bare numbers are used for multiplets in CSresolver and get_iso_statw.
Use of enums (ipDOUBLET, etc), instead, lends to more readable code.

We need to make sure that the fix in connection to the bug with the
order of the 'save species band' and 'init "honly.ini"' commands is
always exercised.

The form of th85rate has a different formula than hmrate. For photoreactions one should use th85rate, not the hmrate. Since, both the th85rate and hmrate  take coefficient values in a similar syntax style (for alpha, beta and gamma), an unintentional mistake happened by me for those reactions.

source/parse_atom_iso.cpp
The SPECIES command still used in some error messages
when the SPECIES command changed to DATABASE

docs/latex/hazy1/optimize.tex
Same mistake in doc for commands with VARY option

The function that computes the statistical weight for levels in He-like
species reports 4n^2 for all levels.  The 4 sums over both singlets and
triplets.  However, the ground has only a singlet state, no triplet.

This is the corollary of an earlier bug, which made the same mistake in
how the n-n' CS is broken up among the fine structure levels of a He-like
ground state.

The implementation omits the exponential term that appears in Eq. (8.30)
of Lebedev & Beigman, as it is folded in its conversion to a collision
strength.  Comment clarifies this point.

Fix bug in H2 line cooling

See merge request !25

Priyanka reported a floating point exception abort in one of her runs.
It came from a straight-up division in CoolH2_GA08.

Enclose all divisions in that function in safe_div calls.

21sep19

See merge request !24

The present coding effectively assumes that both singlets and triplets
are present in the ground of He-like ions, when in reality the ground
is a singlet.

The He-like ground now has a statistical weight of 1.  Fixed.

Hazy claims that the FeII bands are always processed, but this happens
only if a 'save' command has been given.  Move the command that enters
the FeII bands onto the stack of bands to be processed to the top of
SpeciesBandsCreate(), to make sure they are always added.

The 'save species bands' infrastructure uses two classes to store info
for each instance of the command.  The bands file and the species name
are stored as part of the t_save struct, while the data for summing the
band emission are stored in species_pseudo_cont.cpp.  Refactor the code
so that only one class exists that holds as data the parsed info as well
as an object of the class that sums up the emission.

Rename function to getSpeciesGeneric() to highlight that it deals to
user input, and move into said file -- a more natural home than
species_pseudo_cont.cpp and species.h.

Also, modify its interface to be similar to matchGeneric().

Gary sent a bug report that triggered an error with setting up the
"FeII_bands.ini" file in a sim that contained a 'save' command, and
'init "honly.ini"', in that order.

The problem occurred because ParseSave() specified "FeII_bands.ini" for
loading.  Iron was disabled later, so the "Fe+" species was not entered
in the list of species known to the code.  This triggered an abort when
the code tried to set up the FeII bands.

Make sure that the species is active before attempting to set up bands.
This requires that the bands are initialized after the atomic data have
been read in -- the relevant calls have been moved to cloudy().

The initialization of pseudo continua has also been moved to cloudy(),
for symmetry.