C/2001 D1 = 2004 X7 = 2008 S2

C/2001 D1 was found by me on 2002 July 05 in archival C2 images of 2001 February 18 09:45 - 12:30 UTC. I described it as "small comet, visible in many additional images". The comet was confirmed July 06 by Derek Hammer and given the SOHO number 478 on July 11. MPEC 2002-O27 (July 20) announced it as C/2001 D1 with only one additional observation and these orbital elements:

Designation T q e peri node incl L B
C/2001 D1 2001 Feb. 19.01 0.0326 1.0 213.99 173.84 14.75 26.9 -8.2

C/2004 X7 was found by me on 2004 December 08 in C2 images of December 07 08:26 - 10:06 UTC. It was on December 09 confirmed by Karl Battams as SOHO-879. He noticed that he could trace the comet back to 04:26. MPEC 2005-A40 was issued 2005 January 14 with the official designation C/2004 X7 and these orbital elements:

Designation T q e peri node incl L B
C/2004 X7 2004 Dec. 07.71 0.0412 1.0 160.55 180.13 21.34 341.9 +7.0

The orbital elements of C/2004 X7 didn't look similar to those of C/2001 D1 and the apparent track of C/2001 D1 as it would look on December 07 was very different from the track of C/2004 X7.

C/2008 S2 was found by me on 2008 September 18 in C2 images of September 17 19:32 - 21:30. The apparent track was similar to that of C/2004 X7 (as it would appear on September 17) and even not very different from that of C/2001 D1. I measured the positions of the new comet and tried to link them with the positions of C/2004 X7. But the measurements were difficult with very few reference stars found by Astrometrica and the results were poor. I couldn't link the observations.
In the meantime Rob Matson had found additional observations on the other side of the sun from September 18 06:54 - 08:54 UTC. These positions could be well measured and with them I got a first linkage with C/2004 X7 on September 19. The previous perihelion time for this linkage was 2001 Feb 18.59, close to the perihelion time of C/2001 D1. Next, I could link the new comet with C/2001 D1 and then all three comets. The solution showed systematic trends in the residuals but a chance alignment seemed to be improbable. I wrote about my findings to Brian Marsden and included these orbital elements (adding that there was a close approach to Jupiter with 1.3 AU on 2003 October 14):

Designation T q e peri node incl
C/2001 D1 2001 Feb. 18.77 0.0518 0.9787 174.34 163.46 17.95
C/2004 X7 2004 Dec. 07.87 0.0488 0.9799 174.58 163.28 18.79
new comet 2008 Sep. 17.95 0.0485 0.9800 174.60 163.27 18.86

Backward integration led to a previous perihelion of 1997 May 01.62. I began to search those images and Rob Matson joined the search later but we found nothing convincing there.

If the identifications were true, both C/2001 D1 and C/2004 X7 had survived their perihelion passages. I searched for addional images of these comets and found a whole post-perihelic set of them (2001 February 18/19 21:30 - 00:15 UTC) for C/2001 D1 but nothing for C/2004 X7. The new positions of C/2001 D1 lined up well with my new orbit but not with the old orbit (published on MPEC 2002-O27). This could reasonably not be coincidental and meant that the new comet was at least identical with C/2001 D1.

Karl Battams confirmed the new comet as SOHO-1532 and remeasured the 2001 positions and added the new ones. MPEC 2008-S81 came out on September 27 18:26 UTC with Karl's new measurements and orbital elements for C/2001 D1 and the new comet, designated as C/2008 S2. Rather unexpectedly Brian Marsden wrote that the attempts "to link pairs (or all three) of these comets (assuming P = 3.8 or 7.6 years) have not been entirely successful".
The new measurements of the old C/2001 D1 positions using the new star catalogue showed differences of up to 3 arcminutes in both right ascension and declination. When I used Karl's new measurements, my initial solution (given above) gradually converged and the systematic trends in the residuals disappeared. I sent my new orbit determinations to Brian and after a short time the final solution appeared on MPEC 2008-S82 on September 27 22:15 UTC:

Designation T q e peri node incl L B
C/2001 D1 2001 Feb. 18.77 0.0498 0.9796 172.02 165.59 18.78 338.0 +2.6
C/2004 X7 2004 Dec. 07.82 0.0469 0.9807 172.38 165.28 19.77 338.1 +2.6
C/2008 S2 2008 Sep. 17.96 0.0466 0.9808 172.40 165.27 19.84 338.1 +2.6

The close approach to Jupiter was given as "1.29 AU from Jupiter on 2003 Oct. 14".

Using my final solution I found a close approach to Mars with 0.05 AU on 2000 December 12 and a close approach to Earth with 0.18 AU on 1997 March 27 about two months before the perihelion at 1997 May 01.627. Rob Matson tried to find the object in NEAT or DSS images. He wrote to me that the comet just missed the DSS plates of March 1997.

It is very difficult to estimate how bright the object is far from perihelion when it is possibly visible from Earth. The observations extend less to than one day on either side of the perihelia. Karl Battams (IAUC 8986) noted that C/2001 D1 was "very faint and tiny (mag about 8), and was slightly brighter when visible on the far side of the sun". About C/2008 S2 he wrote that it "was tiny and stellar in appearance (mag 7.5-8.0) and appeared even fainter on the far side of the sun". So from this C/2001 D1 and C/2008 S2 appeared at almost the same brightness. But at the perihelion of 2001 the object passed behind the sun, while at perihelion in 2008 it passed in front of the sun, which should have enhanced its brightness considerably (if it has not faded by just the right amount for compensating this enhancement).
No brightness estimate was given for the 2004 apparition (behind the sun with small phase angles), so the evolution of brightness was undetermined. Thus I have measured all images from the three apparitions following the same procedure as for
P/1999 R1 = 2003 R5 = 2007 R5, but I didn't correct the magnitudes for phase angles and got this result:

My measurements for C/2001 D1 (black dots), C/2004 X7 (blue), C/2008 S2 (red).

Unfortunately, the measurements for C/2001 D1 show much scatter, and the magnitudes of C/2001 D1 and C/2008 S2 appear to be a little less bright than Karl Battam's estimates, but the brightnesses of C/2004 X7 and C/2008 S2 (and the centers of the clouds of black dots of C/2001 D1) line up rather well without the corrections for phase angles. This indicates that the dust to gas ratio is very small and/or the dust particles are very small (see the footnote to Marsden comet C/1999 J6 in appendix C of Matthew Knight's thesis).

So perhaps this is not a comet at all? No, an asteroid with mag 8 passing behind the sun should be completely unvisible in SOHO LASCO images when it passes in front of the sun illuminated from behind. What we see here is a very gassy coma with very few or very small dust particles and too small to be resolved by the LASCO C2 instrument with a pixel size of about 12 " (corresponding to ~9000 km at the sun's distance). The perihelion distances of about 0.05 AU are close enough to the sun to evaporate silicates (Sekanina and Chodas 2005). It's our decision to call such objects asteroids with a coma or (nearly) dead comets. The orbits with their very high eccentricities are cometary, so I would still call this a (possibly nearly extinct) comet.

Rainer Kracht, 2009 February 17

The comet of 1997 May 1

When Rob Matson and I had searched the images of 1997, we searched (without success) for an object with at least five positions which are necessary for the confirmation of a SOHO comet.. But even before Rob had joined the search I had found three positions of what looked like a small comet entering the C2 field close to the expected path and with about the right velocity. I had a forth possible postion inside a bright solar streamer, but this was just a spot a bit brighter than the surroundings. After the object had left the streamer, the path was already rather close to the sun with increased background brightness and increased vignetting from the occulter. Later it was probably already too far from perihelion and too faint.

I tried to measure the three positions with Astrometrica; but my first attempt was unsuccessful. It's sometimes very difficult to solve C2 images with Astrometrica. Faint reference stars are indistinguishable from noise. I tried these measurements a second time at the beginning of 2009 April. The residuals of these measurements with the solution of C/2001 D1 = 2004 X7 = 2008 S2 were similar. That was the first sign that the three positions really belong to C/2001 D1 = 2004 X7 = 2008 S2. But how to prove it? Astrometrica had allowed only linear fits with very few reference stars. Blinking the images I found many more stars, so I decided to measure these stars and the comet in all three images and to reduce these measurements afterwards. This is just what Karl Battams and Brian Marsden do: Karl measures the stars and the comet and Brian reduces these measurements. My reductions with a linear Turner solution showed that the three 1997 positions could be linked with C/2001 D1 and 2004 X7.
Brian Marsden kindly reduced my measurements of the 1997 observations and a few days later he got the measurements from Karl Battams and reduced them. Karl and I used different sets of reference stars but the resulting positions agree well..

The four apparitions can be satisfyingly linked with the inclusion of nongravitational parameters. Without these parameters large (about one arcminute) residuals appear for the 1997 observations. But such errors have been seen in the positions of SOHO LASCO C2 comets, resulting from poorly placed reference stars. The reference star closest to the comet is HIP 11672 (TYC 638 491) used in Karl's measurements. This star is only 79, 48, 56 pixel away from the comet in the three images. Errors in the position of the comet of about one arcminute (5 pixel) can be excluded. So we are left with the conclusion that nongravitational forces are acting on this object, which we now can call a real comet.

The new observations received the designation C/1997 J6 (MPEC 2009-H56) and Brian Marsden wrote: "Despite the poor quality of the SOHO observations, a purely gravitational computation from the four apparitions appears to leave significantly systematic residuals".

My measurements for C/1997 J6 (green dots), C/2001 D1 (black), C/2004 X7 (blue), C/2008 S2 (red).

Rainer Kracht, 2009 April 27

The return of 2012 June 30

I have found the comet again in SOHO LASCO C2 images of 2012 June 30. It didn't receive a SOHO number and no measurements were made. Nothing more happened for almost three years, until Alan Hale wrote to me on 2015 May 01: "According to my calculations, the multiple-return SOHO comet C/2001 D1 (I guess it should be "P/2001 D1") passes somewhat close to Earth (0.16 AU) next March while inbound to perihelion.". Perhaps it could be possible to image the comet from the ground in March 2016, but this would need a good ephemeris. So I measured now the positions of the comet in the images where I had found it in 2012. The new observations can be linked to all of the previous observations with the nongravitational parameters from MPEC 2009-H56 (A1 = +0.0002, A2 = -0.0002). The resulting orbital elements are:

Designation T q e peri node incl L B
C/1997 J6 1997 May 01.66 0.0494 0.9797 172.13 165.52 18.81 338.1 +2.5
C/2001 D1 2001 Feb. 18.92 0.0498 0.9796 172.12 165.52 18.75 338.1 +2.5
C/2004 X7 2004 Dec. 08.14 0.0469 0.9807 172.48 165.21 19.74 338.1 +2.5
C/2008 S2 2008 Sep. 18.42 0.0467 0.9808 172.51 165.20 19.81 338.1 +2.5
C/2012 M 2012 July 01.09 0.0467 0.9806 172.51 165.19 19.78 338.1 +2.5
C/2016 2016 Apr. 11.13 0.0456 0.9812 172.64 165.11 20.17 338.2 +2.5

The return of 2016

Closest approach to the earth will be on 2016 March 13 with 24.0 Gm (0.161 au). Only four days earlier the comet will be at elongation 92.5 deg from the sun and could be detected by NEOWISE at a distance of 0.2 au. But it will move at about 6.4 deg/day preventing an automatic detection with WMOPS (WISE Moving Object Processing System), which has a velocity limit of 5 deg/day according to http://arxiv.org/pdf/1406.6025v1.pdf.

Rainer Kracht, 2015 May 03