Detection of weak activity comets

Juan Lacruz, La Caņada Observatory (J87) Oct. 2006

The Full width at half maximum method

When observing weakly active comets on CCD images, the eye may not discern the light distribution of a point source (point spread function) from the distribution of an extended source and many times it is very difficult to tell whether the object is a comet or not.

It often helps considering the full width at half maximum distribution of the punctual objects in the image, comparing the average +- the standard deviation to that of the observed object. When the measured width of the object's light distribution is clearly greater than the average we may be observing a comet

Since these widths are computed by an automated algorithm there're no efects due to the observer appreciation and thus are objective. This method should be understood as an aide and used in combination with other techniques like animations, false colour renditions, isophote maps etc. to provide evidence enough to ascertain the cometary nature

The following graphics show how the FWHM of a typical, weak activity comet, grows compared to that of the sorrounding stars as the photometric aperture grows.

In this case, working at 2.5" per pixel resolution, the curves show that at least a radius of 4 or 5 pixels (10" to 12") should be choosen in order to let enough coma to contaminate the photometric aperture, the effect is measurable as a wider than normal point spread function.

FWHM growth % relative to surrounding stars comet 2005 S3 FWHM growth % relative to surrounding stars comet 2006 M3
In summary the method consist in the following steps :
  1. Stack the images, with Astrometrica, on the body's motion, use a wide enough photometric aperture.
  2. Measure the FWHM and signal to noise ratio of the body
  3. Open the astrometrica log and select the astrometric information, copy and paste it to an excel sheet.
  4. Optionally, discard all the periferic data points (less than 20% of the total image width from the margin)
  5. Plot the data in a dispersion chart

To check the validity of this method I've tried it in all the confirmation of NEOCP object resulting in comets that I did from La Caņada since 2003. A radius of 4 pixels was used for all the astrometry :

The following set of plots and images show statistics of the signal to noise ratio of field stars as a function of the full width at half maximum, the mean value of the FWHM would be a measure of the resulting quality of the observation accounting for the combined effects of seeing, focus, tracking, vibrations etc. A red dot has been plotted representing the S/N and FWHM of the comet, the more to the rigth the more diffuse is the comet. On the rigth column an image of the object is shown for evaluation and comparison purposes.

Non periodic comet C/2002 CE10, in the image no obvious activity is seen, neither the FWHM is clearly different from that of background stars, thus this is a non conclusive observation. A bigger telescope with better resolution at a better seeing place is needed in such cases.
Comet 2003 H3, even when the image doesn't show any evidence, the FWHM plot shows a 8" false nucleus, clearly wider than sorrounding stars. This case shows the power of this method.
P/2003 KV2, non conclusive observation.
P/2003 O3, no evidence of activity in the stack of images and clearly wider psf, FHWM method indicates this could well be a comet.
C/2003 W1, both, evidence in the image and in the FWHM plot, in this case the method supports the new object is a comet.
P/2004 F1, No clear evidence of activity in the stacked image and much wider than sorrounding stars psf. The method supports this one being a comet.
C/2004 L1, too crowded images, no evidence in psf plot neither in stacked image. FWHM does not always help!
C/2004 P1, fuzzy in the image with low S/N ratio, much wider than background stars psf, seems cometary but not enough signal to be sure.
C/2005 N4, non conclusive observation.
C/2005 Q1, obvious activity in the stacked images, also helps the false colour rendition, the psf FWHM is much greater than that of the background stars.
C/2005 Q4, obvious activity in both, image and psf plot.
2005 S3, obvious activity in both, image and psf plot.
P/2006 M3 = P/1889 M1 (Barnard 2), no coma seen in the stacked image and much wider psf than sorrounding stars in the psf plot.

In summary :

Total observations13
Evidence in the image and in the FWHM distribution430%
No evident in image, evident in the FWHM distribution4 30%
No evidence in image, no evidence in psf plot5

The FWHM method was of help in 60% of the considered cases.