Hi Charles,

It's taken me almost a month to reply but here are some comments on the
way that comphot calculates the coma diameter. I note that you say that
you are sometimes seeing reported coma diameters that are too large.
That is odd since most of the time comphot gives a diameter that is too
small since the sky background is not flat. If you have any specific
examples of that I can chase them up.

The attached shows some results from an observation of C/2020 R4 on
April 17 taken in good conditions with the sky about as dark as it ever
gets from here.

The frames used are sigma-clip stacks of 31x60s subframes which are then
plate solved against Gaia DR2. These are 15x15 cut-outs from my 33x22
arcmin field.

The magnitude zero point (25.96) is derived from ensemble photometry
against around 100 Gaia G magnitudes. The sky background based on this
ZP is 18.97 mag/arcsec^2.

Comphot attempts to calculate the coma diameter by estimating the RMS
sky noise and then moving out in circular annuli until the coma
brightness drops to below 1 sigma. In this case the 1-sigma threshold is
5.5 ADU and the sky background is 1035 ADU so the coma cut-off is 0.5%
of the sky level. That kind of sensitivity requires very close attention
to cal frames, particularly flat fields.

In the attached you can see the estimate that Comphot makes using this
method (top right) compared with a manual hard linear stretch overlaid
with the comphot result (bottom right). The estimate in this case
appears to be about right at 190 arcsec radius. As you say, observers
should check this as a matter of routine to make sure that the diameter
comphot computes looks reasonable.

Comphot does output a log of the intermediate estimates so that you can
plot a coma profile or a magnitude growth curve as I have done here.
This is the long list of numbers in the middle of the log output after
the "Total pixels requested but outside frame" line. I regularly plot
these just to check that things look sensible. Check also the "skycheck"
plot which will show any non-uniform sky background.

The magnitude is calculated by integrating the pixel counts within the
190 arcsec radius assuminhg that the coma is circularly symmetric around
the photocentre. This should avoid any contribution from the tail. The
sky background is taken from the sigma-clipped mean of the outer annulus
shown on the plot.

The results of this analysis for April 17.1 are a magnitude of 10.6 and
a detectable coma diameter of 6.3 arcmin. Looking at results on COBS my
coma diameter is a bit on the large side (but I think reasonable) and
the magnitude is probably 0.5 fainter than most other people are getting.

You are certainly correct that for comphot to work correctly the
background sky needs to be as flat as possible. That is the case for any
observation where you are trying to detect something that is 1% or less
of the sky brightness. I think this is the main area where people are
having problems. If you are willing to share any FITS files where you
have made observations that show an abnormally large coma I would be
happy to look at them to try to work out what is happening.

Nick.

On 20/03/2021 00:23, Charles S Morris - cometguy3783 at yahoo.com (via
baa-comet list) wrote:
>   A few general comments -
> Visual
> 1) Elevation was not given in the original ICQ format because there was no obvious correlation with the reported brightness estimates (at that time) and I haven't seen a statistical study that proves there is a significant correlation.  I have made many visual observations at 1500m with a limiting stellar  NE magnitude of 7.4.  I have also made observations closer to the city at 1000m (still good sky conditions, but not as good) and some from my house at 150m. I even observed Hale-Bopp on one night at 4,200m in the darkest sky you can imagine.  And I don't see a significant difference in brightness estimates.  This is not to say that a significant elevation doesn't help with things like extinction, but it is not the main factor.  {Elevation was part of the IHW database, as I recall]
> 2) A measure of sky brightness was not included in the original ICQ format because there was no standard measurement for that at the time and we were limited to the 80 characters (punch cards).  For visual observing, measured coma diameter was thought to be a good indicator of conditions.  I do not believe an analyst should use sky brightness as selection factor in choosing observations to analyze.
> 3) The primary impact on the reliability of visual observations was found to be observer experience based on a study with the IHW Halley database by Steve Edberg and myself.  I still believe this to be true today.  All visual observations are not equal.  Experienced observers can get reasonable results even in compromised situations.  Their eyes-brain have been trained to observe comets.
> By the way, my definition of "an experienced visual observer" is someone who has observed 30-50 comets with at least 10 observations per comet.  This can be obtained in 3-5 years of serious observing.  Too often observers only observe the bright comets and can not be bothered when the comets are fainter than 6 mag..  With Halley (which was visible for more than a year), requiring 10 observations for inclusion in the light curve substantially reduced the scatter in the visual light curve.
> In 1977, five experienced observers including myself, John Bortle, Steve O'Meara slipped away from an AAVSO meeting to observe Comet Kohler.  Using the same comparison stars, binoculars (7x50B or 10x50B) the estimated brightness estimates were within +-0.1 magnitudes of each other.  I take that as the goal of any visual light curve.
> I will note my own personal experience has been that gassy comets are easier to observe through moonlight and city lights.  Obviously there is a spectral difference with dusty comets, but the exact reason for this is unclear.  Not sure if my eyes are more blue sensitive or what the reason is..
> CCD
> With no offense to Nick James or COMPHOT, I take such program determined coma diameters with a grain of salt.  I have seen a number of examples in COBS  where COMPHOT produced exaggerated coma diameters with corresponding bright magnitudes.  In one case the reported coma diameter was at least 5 times what was being reported by other CCD observers, including visual observers.  [By the way -Agreeing with visual (eye) magnitude observations is not a justified reason to accept an unreasonable coma diameter estimate.]  There are real differences between CCD and visual (eye) observations.  Smaller aperture CCD systems are now giving magnitudes closer to visual (eye) observations.
> With my own experiments with COMPHOT I have generated significantly too small of a coma diameter and significantly too large of a coma diameter based on actual transects through the coma.  The REASON for this is probably that the field of the image is not perfectly flat.  Nick James emphasizes the need for a flat field.  How many observers get a totally flat field?  My guess is relatively few.  I admit that between the moon, twilight, city lights in some directions my fields are not totally flat although I do use a flatting routine where it is necessary.  But the fields are rarely as flat as Nick would recommend.  Unless your field is totally flat, a coma diameter based on COMPHOT maybe suspect.
> I measure coma diameter using a transect through the coma.  Then I set the r parameter in COMPHOT.  In my opinion, every CCD observer should do that to confirm their coma diameter.  For the record, there is NO hidden coma in the background noise that can be magically pulled out.  [If there is, please prove it to me.]  I suspect in some fields there are a lot of faint stars that can be included in the coma (even with the star removal routine) which can cause the coma to grow larger than it should.
> If you have great confidence in COMPHOT determining coma diameters, you should still put a transect through the coma (at right angles to the tail) to verify that your result is reasonable.
> By the way - magnitudes, using Astrometrica, even with a field with a really bad gradient affects the zero magnitude by less than 0.05 magnitudes.  I tested that.
> Look, I am not saying that COMPHOT is a bad program, it gives reasonable brightness results when I specify the diameter.  But observers should be aware that it should not be used as a black box.  What's the saying?  Trust, but verify.  That should be the case with any photometry program.
> I do wish COMPHOT had a graphic output showing the coma brightness as a function of diameter.  A lot issues, if they exist, can be seen in such a plot.
> Charles S. MorrisAssociate Editor, ICQDreamweaver ObservatoryFillmore, CA USA
>
>
> .
>      On Friday, March 19, 2021, 12:33:03 PM PDT, jjgonzalez jjgonzalez <jjgonzalez@telecable.es> wrote:
>
>   Peter, Thomas abd friends,
>
> Many thanks for the supplementary information and data.
>
> Up to now we have these COBS recent estimates for C/2020 R4 (ATLAS) :
>
> Visual (7) :
> Mar. 17.19, 9.6, 3' (C. Labordena, Culla, 1000 m, Spain, 0.2-m Schmidt-Cassegrain)
> Mar. 16.31, 10.0, 2' (J. de Souza Aguiar, Campinas, Brasil, 0.27-m Newtonian reflector)
> Mar. 15.79, 9.9, 3' (M. Mattiazzo, Swan Hill, near sea level,  Victoria, Australia, 25x100 binoculars)
> Mar. 15.30, 10.0, -- (J. de Souza Aguiar,  0.27-m Newtonian reflector)
> Mar. 14.31, 9.8, -- (J. de Souza Aguiar, 0.27-m Newtonian reflector)
> Mar. 13.82, 10.4, 2' (P. Camilleri, Australia, 0.406-m Newtonian reflector)
> Mar. 13.21, 8.4, 6' (J. J. Gonzalez Suarez, Alto del Castro - Aralla, 1720 m,  Spain, 0.203-m Schmidt-Cassegrain)
>
> CCD - CMOS (3) :
> Mar. 17.21, 10.1, 4.7' (P. Carson, remote, Fregenal de la Sierra Spain, 580 m, Spain, 0.315-m Cassegrain reflector)
> Mar. 15.15, 9.9, 10' (T. Lehmann, remote, Hakos, 1830 m, Namibia, 0.2-m Newtonian reflector; 0.26-deg tail in p.a. 265)
> Mar. 09.78, 10.7, 1.5' (M. Mobberley, remote, Siding Spring, 1160 m, Australia, 0.5-m Ritchey-Chretien)
>
> When comparing these two lists, there is an interesting result :
> most estimates ( 6 visual, 2 CCD - CMOS ) are close to m ~ 10.0 ...
> ... but they correspond to clearly different diameters !  --> ( 2' - 3' visual, 4.7' - 10' CCD - CMOS ).
> This is somewhat peculiar ...
>
> Though I have a master's degree in Astrophysics ( Universidad Complutense, Madrid, 1970 - 1976 ), I am not a photometry theorist ... only an old style visual comet observer, but  ,,,
>>From my long observing experience, I know that for a detailed comparison of comet estimates data ( m1, coma diameter, DC, tail length if present ) and drawing lightcurves, in relation with the conditions of the observing site, we must know at least two factors not included in the "classic" abbreviated comet report formats :
> - Night sky brightness.
> - Elevation of the observing site.
> These factors are especially important for diffuse comets with faint outer coma, located at low altitude over the horizon.
>
> My estimate for C/2020 R4 ( m1=8.4, Dia.=6', DC=3 ) is consistent with the wider observed diameter from a dark location at 1700 m in the Cantabrian Mountains.
>
> As Peter say, lets hope my prediction of a m1 ~ 7.0 mag comes to be ...
>
> Best regards and clear skies,
>
> J. J. Gonzalez Suarez
>
> P.S.: Disclaimer : I'm not an specialist on comet magnitude predictions ...
>
> -------------------------------------------------------------------------------------
>
>
>
>> On 18 March 2021 Thomas Lehmann wrote :
>>
>> I have observed the comet a few days ago remotely under pristine skies of Namibia
>> using a 20cm telescope, with CMOS camera and green filter. I measured the total
>> brightness within a large aperture of 10'
>>    2021-03-15.15 UT, m1=9.9, coma dia. 10'
>>
>> Thomas
>>
> ------------------------------------------------------------
>
>> On 18 March 2021 Peter Carson wrote :
>>
>> Hello Juan Jose,
>>
>> My measurement of C/2020R4 was made from Fregenal de la Sierra in Extremadura at an altitude of 560m under skies that were SQM 21.8 at the zenith with low humidity. My image records a coma dia of 5’. The Comphot photometry tool used a measuring aperture of 4.7’ so measured virtually all the apparent coma. I’m confident my result is consistent with all my other photometry in COBS.
>>
>> Lets hope your prediction of a m1 7.0 mag comes to be.
>>
>> All the best
>>
>> Peter
>>
> ----------------------------------------------------------------
>> On 18 March 2021 J. J. Gonzalez wrote :
>>
>> Nick, Peter, and friends,
>>
>>      In my condition of "bright outlier", I kindly disagree when you say "It probably won't get much brighter through March and April".
>>
>>      COBS recent visual estimates around m1~10.0 are associated to a observed coma diameter in the 2' - 3' range.
>>
>>      From dark mountain skies the comet shows a wider diffuse outer coma. In my recent observation :
>>
>> C/2020 R4 (ATLAS):
>> 2021 Mar. 13.21 UT: m1=8.4, Dia.=6', DC=3, 20 cm SCT (77x).
>> [ Altitude: 12 deg. Mountain location, very clear sky.
>> Sidgwick method. Tycho-2 comparison stars. SQM: 21.0.].
>> ( Alto del Castro, Leon, Spain, alt. 1720 m; SQM 21.5 at zenith ).
>>
>> This estimate is in good agreement with the formula :
>>
>> m1 =  7.5 + 5 log delta + 10.0 log r
>>
>> providing m1~7.0 at Earth's close approach ( delta = 0.46 AU on 2021 Apr. 23 ), located high in the sky on the Hercules–Corona Borealis border.
>>
>>      It will be an interesting photometric evolution to follow ... weather permitting.
>>
>> Best regards and clear skies,
>>
>> J. J. Gonzalez Suarez
>
> ----------------------------------------------------------------------
>>
>>> On 18 March 2021 Nick James <ndj@nickdjames.com> wrote :
>>>
>>>
>>> Peter,
>>>
>>> Thanks for the image obtained under difficult circumstances.
>>>
>>> Martin Mobberley got 10.7 on March 9.78 and visual estimates on COBS are
>>> all around 10.0 now with one bright outlier. It probably won't get much
>>> brighter through March and April. At least it will become easier for us
>>> to observe from the UK over the next few weeks.
>>>
>>> Nick.
>>>
>>> -----------------------------------------------------------------------------------
> To unsubscribe from this list please go to http://archives.simplelists.com
> To unsubscribe from this list please go to http://www.simplelists.com/confirm.php?u=Aj4vCBHI6JSRpYRB2DWvMTYlWL2vuAnC
>