line convergence at outer boundary of large column density clouds (trac #327)

[cloudy-bot]

reported by: @CloudyLex

The PDR sims in the test suite include the DOUBLE command for the reasons described in the Quick Start Guide, Section 8.3.3 on "heads up" for PDRs. There are two reasons why. First, the models stop at Av = 10, the convention for PDR models. Real PDRs do not stop but extend into an unmodeled background molecular cloud. That unmodeled material is approximated when the DOUBLE command is included.

If DOUBLE is not included then a PDR will be very optically thick at the shielded face, with most coolants affected by line transfer. As the outer edge is reached the lines will become optically thin, on the very surface of the shielded face. For a line like Lya the total optical depth might be 1e8 and the line photons become freely escaping when optical depths within a few are reached at the shielded face. Conditions will change greatly at the outer skin due to this. We don't now have logic to detect this condition and adjust the zoning to resolve this thin skin at the shielded face, because this does not occur when a PDR is set up with the DOUBLE command. Physically, it does not happen in nature, for a PDR, because of the unmodeled molecular cloud.

The PDR sim pdr_coolbb was added to exercise the code in an extreme radiative case. I forgot to include the DOUBLE command. As a result ITERATE TO CONVERGENCE failed and the answer depends on circumstances with lots of botches coming and going. This was due to an oversight on my part and I have now added the DOUBLE command and ITERATE TO CONVERGENCE.

This ticket is because it would be good if we could converge such models. It will take some development of the zoning logic for this circumstance. To exercise the problem comment out the DOUBLE command in the pdr_coolbb sim

Migrated from https://www.nublado.org/ticket/327

{
    "status": "new",
    "changetime": "2019-02-04T13:07:06Z",
    "_ts": "1549285626580696",
    "description": "The PDR sims in the test suite include the DOUBLE command for the reasons described in the Quick Start Guide, Section 8.3.3 on \"heads up\" for PDRs.  There are two reasons why.  First, the models stop at Av = 10, the convention for PDR models.  Real PDRs do not stop but extend into an unmodeled background molecular cloud.  That unmodeled material is approximated when the DOUBLE command is included.\n\nIf DOUBLE is not included then a PDR will be very optically thick at the shielded face, with most coolants affected by line transfer.  As the outer edge is reached the lines will become optically thin, on the very surface of the shielded face.  For a line like Lya the total optical depth might be 1e8 and the line photons become freely escaping when optical depths within a few are reached at the shielded face.  Conditions will change greatly at the outer skin due to this.  We don't now have logic to detect this condition and adjust the zoning to resolve this thin skin at the shielded face, because this does not occur when a PDR is set up with the DOUBLE command.  Physically, it does not happen in nature, for a PDR, because of the unmodeled molecular cloud.\n\nThe PDR sim pdr_coolbb was added to exercise the code in an extreme radiative case.  I forgot to include the DOUBLE command.  As a result ITERATE TO CONVERGENCE failed and the answer depends on circumstances with lots of botches coming and going.  This was due to an oversight on my part and I have now added the DOUBLE command and  ITERATE TO CONVERGENCE.\n\nThis ticket is because it would be good if we could converge such models.  It will take some development of the zoning logic for this circumstance.  To exercise the problem comment out the DOUBLE command in the pdr_coolbb sim",
    "reporter": "gary",
    "cc": "",
    "resolution": "",
    "time": "2015-06-03T19:00:13Z",
    "component": "radiative transfer",
    "summary": "line convergence at outer boundary of large column density clouds",
    "priority": "good to do",
    "keywords": "",
    "version": "trunk",
    "milestone": "no milestone",
    "owner": "nobody",
    "type": "defect - convergence"
}

[cloudy-bot]

@CloudyLex commented:

change to pdr_coolbb committed in trunk r10441.

[cloudy-bot]

@peter-van-hoof-noaccount changed milestone from "" to "no milestone"