Fixed bug in which qdisk structure was incorrectly zeroed before photon ...

[jhmatthews]

...generation.

This fixes a silly error in which the heating stored in the qdisk structure was zeroed before it was even used, meaning that the SEDs used for the ionization state was identical if you were using disk illumination mode 2 or mode 0. To fix, I simply move the zeroing line after photon_gen in python.c. I've tested that this fix moves the illuminating SED up to roughly the expected level.

Interestingly in the spectral cycles the SEDs were different because it still had the information from the last ion cycle.

I just looked in python 58 and the same bug appears to be present there. I'm now a little worried that this may have affected @ssim 's 2005 paper, as a number of reprocessing models were used in that study.

[ssim]

HI James,

If I recall correctly, in the 2005 paper we used an analytic formula for the amount of reprocessing, which I think assumed that the disk is directly illuminated by the central star (I'd have to check the formula) - but I don't think we used the option to try and have Python predict how much re-radiation we expected. So I think we used reprocessing options "1" (which is the reflecting case) and "3" which was this analytic approximation. I don't think either of those cases should be affected by what is in qdisk (but would be good to check).

Stuart

On 13 Feb 2014, at 18:43, James Matthews [email protected] wrote:

...generation.

This fixes a silly error in which the heating stored in the qdisk structure was zeroed before it was even used, meaning that the SEDs used for the ionization state was identical if you were using disk illumination mode 2 or mode 0. To fix, I simply move the zeroing line after photon_gen in python.c. I've tested that this fix moves the illuminating SED up to roughly the expected level.

Interestingly in the spectral cycles the SEDs were different because it still had the information from the last ion cycle.

I just looked in python 58 and the same bug appears to be present there. I'm now a little worried that this may have affected @ssim 's 2005 paper, as a number of reprocessing models were used in that study.

You can merge this Pull Request by running

git pull https://github.com/jhmatthews/python qdisk
Or view, comment on, or merge it at:

https://github.com/agnwinds/python/pull/69

Commit Summary

• Fixed bug in which qdisk structure was incorrectly zeroed before photon generation
File Changes

• M python.c (28)
Patch Links:

• https://github.com/agnwinds/python/pull/69.patch
• https://github.com/agnwinds/python/pull/69.diff
—
Reply to this email directly or view it on GitHub.

[jhmatthews]

I've just checked Stuart and you're absolutely right. Reprocessing options 1 and 3 are never affected by this, just 2 and I don't think we've used this for anything, which is lucky!

Merging this right away.