Results from the Worldwide Coma Morphology
Campaign for Comet ISON (C/2012 S1)1
Nalin H. Samarasinha & Beatrice E.A. Mueller (Planetary Science Institute, USA),
Matthew M. Knight (Lowell Obs, USA), Tony L. Farnham (Univ of Maryland, College
Park, USA), John Briol (Spirit Marsh Obs, USA), Noah Brosch (Wise Obs, Tel Aviv Univ,
Israel), John Caruso (Temecula, CA, USA), Xing Gao (No 1 Senior High School,
Urumqi, China), Edward Gomez & Tim Lister (Las Cumbres Observatory Global
Telescope Network, USA), Carl Hergenrother (Univ of Arizona, USA), Susan Hoban &
Roy Prouty (Univ of Maryland, Baltimore County, USA), Mike Holloway (Holloway
Comet Obs, USA), Nick Howes & Ernesto Guido (Remanzacco Obs, Italy), Man-To Hui
(Univ of California, Los Angeles, USA), Joseph H. Jones, Tyler B. Penland, Samuel R.
Thomas & Jim Wyrosdick (Univ of North Georgia, USA), Nikolai Kiselev & Aleksandra
V. Ivanova (Main Astronomical Obs, National Academy of Sciences, Ukraine), Thomas
G. Kaye (Raemor Vista Obs, USA), Jean-Baptist Kikwaya Eluo (Vatican Obs, USA),
Betty P.S. Lau (Hong Kong), Zhong-Yi Lin (National Central Univ, Taiwan), José Luis
Martin (Carpe Noctem Obs, Spain), Alexander S. Moskvitin (Special Astrophysical Obs,
Russian Academy of Sciences, Russia), Martino Nicolini (Cavezzo Obs, Italy), Brian D.
Ottum (Saline, MI, USA), Chris Pruzenski, David C. Vogel, Leo Kellett, Valerie Rapson,
Joel Schmid, Brandon Doyle, Frank Dimino & Stephanie Carlino (Astronomy Section,
Rochester Academy of Science, USA), Margarita Safonova, Jayant Murthy & Firoza
Sutaria (Indian Institute of Astrophysics, India), David G. Schleicher (Lowell Obs, USA),
Colin Snodgrass (The Open University, UK), Cihan T. Tezcan & Onur Yorukoglu
(Ankara Univ, Kreiken Obs, Turkey), David Trowbridge (Tinyblue Obs, USA), Dennis
Whitmer (Eagle Nest Digital Obs, USA), Quan-Zhi Ye (Univ of Western Ontario,
Corresponding Author information:
Nalin H. Samarasinha
Planetary Science Institute
1700 E Ft Lowell Road, Suite 106, Tucson, AZ 85719, USA
(Tel: 1-520-547-3952, Email: email@example.com)
Submitted to Planetary and Space Science (March 04, 2015)
Revised: September 27, 2015
1 We dedicate this paper to our colleague and friend Martino Nicolini who on the 29th January 2015
passed away after a long fight against cancer. Martino, who was a nuclear engineer by profession, was
an avid amateur astronomer and he collaborated with us as well as with many other organizations. He
epitomized what the professional-amateur collaborations could accomplish and will be sorely missed by
We present the results of a global coma morphology campaign for comet C/2012 S1
(ISON), which was organized to involve both professional and amateur observers. In
response to the campaign, many hundreds of images, from nearly two dozen groups
were collected. Images were taken primarily in the continuum, which help to
characterize the behavior of dust in the coma of comet ISON. The campaign received
images from January 12 through November 22, 2013 (an interval over which the
heliocentric distance decreased from 5.1 AU to 0.35 AU), allowing monitoring of the
long-term evolution of coma morphology during comet ISON’s pre-perihelion leg. Data
were contributed by observers spread around the world, resulting in particularly good
temporal coverage during November when comet ISON was brightest but its visibility
was limited from any one location due to the small solar elongation. We analyze the
northwestern sunward continuum coma feature observed in comet ISON during the first
half of 2013, finding that it was likely present from at least February through May and
did not show variations on diurnal time scales. From these images we constrain the
grain velocities to ~10 m s-1, and we find that the grains spent 2-4 weeks in the sunward
side prior to merging with the dust tail. We present a rationale for the lack of continuum
coma features from September until mid-November 2013, determining that if the feature
from the first half of 2013 was present, it was likely too small to be clearly detected. We
also analyze the continuum coma morphology observed subsequent to the November
12 outburst, and constrain the first appearance of new features in the continuum to later
than November 13.99 UT.
Comet C/2012 S1 (ISON) (hereafter comet ISON) was discovered on September 21,
2012 at a heliocentric distance, rh, of 6.3 AU (Novski and Novichonok 2012). Soon it
attracted worldwide interest because of its extremely small perihelion distance
(0.0125 AU = 2.7 solar radii) and the prediction that it may become a bright naked eye
object based on its brightness behavior soon after discovery. The discovery of comet
ISON at a large heliocentric distance and the favorable observing geometry during most
of the apparition, partially facilitated by the comparatively high orbital inclination (61.°9),
made it the first instance that the behavior of cometary activity of a sungrazer2 was
monitored for a large range of heliocentric distances. This range included the large
distances (i.e., > 2-3 AU) where the nuclear activity is dominated by super volatiles
(e.g., CO, CO2), the region where water is the primary driver of activity, as well as the
extremely small heliocentric distances when the sublimation of refractory material,
including metallic species that are present in the dust grains, could occur. In other
words, the heliocentric distances covered a range over which the incident solar flux
increased by a factor of about 2.5x105 making this the first time that the effects due to
such a large range of solar fluxes on the nucleus were monitored for any comet.
Anticipating these extreme conditions, there were many predictions on what might
happen to the nucleus as it approached perihelion. The predictions included those
pertaining to cometary activity and nuclear rotational state, as well as the ultimate fate
of the nucleus (e.g., Samarasinha and Mueller 2013, Knight and Walsh 2013, Ferrin
Past work on morphological studies of cometary comae has demonstrated the value for
inferring properties of the nucleus such as rotation period, seasonal activity changes,
and pole orientation (e.g., Schleicher et al. 2003, Farnham et al. 2007, Farnham 2009,
Knight et al. 2012 and references therein). With the goal of temporal monitoring of
comet ISON’s coma features, the campaign organizers3 coordinated a Worldwide
Campaign of Coma Morphology (see http://www.psi.edu/ison for the call for images).
This effort was also carried out in coordination with the NASA Comet ISON Observing
Campaign (CIOC) but was independent of the CIOC activities (see
http://www.isoncampaign.org for additional information on CIOC). After May 2013,
comet ISON was not observable for more than a few hours from any geographic
location due to the relatively small solar elongation, so a global effort, incorporating
images from both professionals and amateurs, was necessary to carry out a detailed
analysis of the coma morphology and its temporal and spatial evolution. Although the
relatively large geocentric distances to the comet during the apparition (>0.85 AU during
the pre-perihelion leg and >0.43 AU during the post-perihelion leg) would not provide
the ideal circumstances needed for producing high-resolution detailed coma features,
the large range of heliocentric distances over which comet ISON might be studied and
the resulting possibility of observing a variety of interesting coma morphologies
motivated us to organize this coordinated worldwide campaign. The campaign solicited
2 A formal definition for the sungrazer comets is provided in Knight and Walsh (2013) in terms of the
3 The campaign was organized by N.H. Samarasinha, B.E.A. Mueller, M.M. Knight, and T.L. Farnham.
both continuum (dust) images as well as gas images of the near-nucleus region of the
comet. We are happy to note that a large number of images were collected from
observers spread around the world at different longitudes.
Monitoring of comet ISON was encouraged for both before and after its perihelion
passage, with an emphasis on the time around perigee (December 26, 2013) when it
was expected to experience rapid changes and would have been observable for many
hours per night in the northern hemisphere. However, the comet started to disintegrate
just prior to its perihelion passage (cf. Knight and Battams 2014, Sekanina and Kracht
2014) leaving us with images only from the pre-perihelion leg.
In Section 2 of this paper, we provide a description of the images collected by the
campaign. In Sections 3 and 4, we discuss the analyses of these images during the pre-
water dominated phase and the water-dominated phase, respectively. Section 5
provides the summary and conclusions of this paper.
Images were collected from both amateur and professional observers, with diverse
ranges of telescope apertures, observing conditions, filters, and data collection
methodologies. These data are far from uniform, but such an approach was necessary
in order to obtain as much longitudinal coverage as possible. Observers were asked to
reduce (remove bias and correct for flat fielding) their own images prior to submitting
them to the campaign. The campaign organizers, using these reduced images, carried
out all image enhancements. The relevant enhancement techniques are described in
Samarasinha and Larson (2014). The coma images we received cover a large interval
from January 12, 2013 (rh~5.14 AU) to November 22, 2013 (rh~ 0.35 AU). Table 1 lists a
summary of information on the images provided by different groups of observers
ordered based on the geographic longitude of the observations. Figure 1 shows the
observational circumstances for all the images received by the campaign. The symbols
for different parameters listed in Table 1 and Figure 1 are identified in the legend of
Figure 1. Most observations listed in Table 1 were not published elsewhere and are
presented here for the first time. Observers submitted different sized images; however,
the pixel scales for all the images were smaller than the respective astronomical seeing.
As can be seen from Table 1 and Figure 1, the majority of the images were obtained
between September and mid-November, 2013, when the comet was bright and the
solar elongation was >30˚. However, as we will discuss in detail in Section 4, the
images from this time interval showed no clearly identifiable coma features in the
continuum until the outburst that started around November 12 (e.g., Opitom et al.
2013a). In contrast, when the comet’s solar elongation was favorable early in the
apparition (i.e., before June, 2013), there were many clearly identifiable coma features
in the continuum. In images of the gas coma, features appeared by November 1, nearly
a month prior to the perihelion of November 28.779 UT (e.g., Opitom et al. 2013b,
Knight and Schleicher 2015). In the following sections, we analyze the continuum
images and provide a rationale for the morphological behavior.
3. Results from the Pre-Water-Dominated Phase
The pre-water-dominated phase (i.e., when the surface temperature of a comet is too
low for water to be the dominant or the most productive volatile, where super-volatiles
such as CO or CO2 could be the primary driver of cometary activity) occurs when a
comet is further than about 2-3 AU from the Sun (e.g., Bockelée-Morvan et al. 2004).
The time when comet ISON was at small solar elongations (i.e., approximately < 30°
from June through August 2013; also Figure 1) coincides with the transition to water-
dominated gas production, providing a natural boundary for the observations. Therefore,
we designate observations prior to July 2013 (i.e., rh>~3.1 AU) as representing the pre-
water-dominated phase, while images after August 2013 (i.e., rh<~2.2 AU) as coinciding
with the water-dominated phase. In this section, we analyze and discuss the coma
morphology and its evolution in the pre-water dominated phase.
Despite the fact that the comet was discovered at 6.3 AU from the Sun, the large
geocentric distance coupled with the low surface brightness of the coma prevented the
acquisition of high signal-to-noise (S/N) images with good spatial resolution. In addition,
viewing of the impact of radiation pressure is the combined effect of the large
heliocentric distance and the small solar phase angle (<10° until late-February with a
minimum of 1.8° on January 11). This impeded any immediate detection of
unambiguous coma morphology other than the dust tail in the earliest images
Figure 1. Behavior of
heliocentric distance rh,
geocentric distance Δ,
solar phase angle
solar elongation (Sun-
and position angle PA
of the skyplane-
projected Sun direction
PA (measured from
north through east) as a
function of time. The
symbols depict the UT
dates when observa-
tions are available from
this global campaign.
The parameters corres-
ponding to each symbol
are identified in the
(Figure 2). However, as the comet moved towards the inner solar system it underwent a
rapid increase in its brightness prior to 5 AU (e.g., Figure 2 of Meech et al. 2013) that
provided the first realistic opportunity for coma morphological studies.
3.1 Sunward Feature in the Continuum Images
The first announcement of the presence of a coma feature in the continuum was based
on enhanced Hubble Space Telescope (HST) images taken on April 10, 2013 (Li et al.
2013). The HST images showed a northwest sunward feature starting at a westerly
direction and curving towards north and then merging with the tail due to radiation
pressure. Then, Howes and Guido (2013) announced the presence of the same
sunward feature in ground-based images taken in early-May (rh~3.9 AU) from the 2m
Liverpool Telescope (LT). Figure 3 shows the entire set of the LT images from this
epoch indicating that there are no clearly discernible changes in the morphology on a
daily timescale. Image enhancement of comet ISON images taken at the 4.3m
Discovery Channel Telescope (DCT) by Knight and Schleicher (2015) provides clear
evidence that the same sunward feature was present from March through mid-May, and
HST observations by Hines et al. (2014) in early May confirm that it was relatively
Figure 2. Inner coma region of two R-band images taken at the Vatican Advanced
Technology Telescope in mid-January (rh~5 AU). The images are enhanced using
division by an azimuthal median technique (e.g., Samarasinha and Larson 2014).
Orange represents brighter regions while black denotes the dimmer regions. Each
panel is approximately 48,000 km across with the nucleus located at the center. The
skyplane-projected PA of the Sun direction was rapidly moving clockwise from 343°
to 320° (respective values are for January 15 and 20). The bright feature in the
southeast quadrant is attributed to dust emitted from the nucleus that was
subsequently swept away in the respective anti-sunward directions over a few
weeks. Low spatial resolution and S/N due to the large geocentric distance prevent
unambiguous detection of coma features close to the nucleus.
unchanged since April even at higher spatial scales. Images of comet ISON taken at the
1.8m Vatican Advanced Technology Telescope (VATT) by Carl Hergenrother in mid-
February show the same sunward feature4 albeit at a lower S/N. Therefore, we infer that
this sunward feature is present at least in the images spanning an interval from mid-
February to mid-May 2013. As June approached, the decreasing solar elongation
prevented continued monitoring of this feature5. Figure 4 shows the temporal evolution
of the feature in the February-May time frame.
For activity originating from a fixed source region away from the poles, one expects to
see morphological variations on rotational timescales. No photometric variations in the
brightness of comet ISON were detected in January by the Deep Impact spacecraft
(Farnham et al. 2013), by the HST in April (Li et al. 2013), or by the Spitzer Space
Telescope in June (Lisse et al. 2013) suggesting any rotational variation should be
small. Despite that, observations taken on November 1, 2013 by the HST, when the
comet was much closer to us (rh=1.00 AU, Δ=1.23 AU), show a single-peak photometric
variation of nearly a factor two with a period of ~10.4 hours which was attributed to the
rotation of the nucleus (Lamy et al. 2014).
4 Generally, we are extremely reluctant to trust the reality of the sunward feature present in these mid-
February images (and to an extent in some March images), as the spatial location of the feature is
extremely sensitive to a single pixel offset of the nucleus position. However, the fact that the PA of the
feature and the general morphology are consistent with subsequent April through May images provide
confidence that the feature is indeed real.
5 Knight and Schleicher (2015) have images from June 11; however, these images were taken at
extremely high airmass and have S/N too low to detect coma features.
Figure 3. The near nucleus region of R-band images taken at the Liverpool
Telescope on May 1, 2, 3, 5, and 7 (left to right). Images are enhanced using the
division by azimuthal average technique, which is nearly identical to the division by
azimuthal median (e.g., Samarasinha and Larson 2014). Each panel is approximately
25,000 km across with the nucleus located at the center. The skyplane-projected PA
of the Sun direction is 270°. The sunward feature originates in an approximately
westerly direction (to the right) and then moves northward and finally merges with the
dust tail which is to the left (East) as a consequence of the radiation pressure effects.
No clearly discernable temporal variations of the sunward feature are detected at this
spatial resolution and scale. The minor changes from one panel to another are due to
low brightness of the comet at this heliocentric distance.
However, no variations in the morphology were detected at rotational timescales either
based on the ground-based images or on the HST images. Due to this fact, Li et al.
(2013) attribute the origin of the sunward feature observed by HST in April to a source
region near comet ISON’s rotational pole. Another possibility is that this feature is due to
the cumulative grain outflow from the sunward side of the nucleus as a response to
insolation and not necessarily from a fixed source region on the nucleus (cf. Belton
3.2 Characterization of Grains in the Sunward Feature
The measured sunward extension of the feature projected onto the skyplane, d, is 4×103
– 6×103 km. Therefore, for this d, V2/
>10-3 km2 s-2 where V is the outflow velocity of
is the radiation pressure parameter (cf. equation (5) of Mueller et al. 2013).
For dominant micron-size grains (i.e.,
~ 0.1; Burns et al. 1979), V should be ~10 m s-1.
Grain outflow velocities for micron sized grains that are of the order of ten or a few tens
Figure 4. Images from February to May (rh decreasing from ~4.8 AU to ~3.7 AU)
showing the evolution of the sunward feature. The date and telescope are given at
the top of each panel, and the corresponding geometric circumstances can be
determined from Figure 1 and Table 1. All images are enhanced with division by
azimuthal median. Each panel is approximately 25,000 km across. The PA of the
Sun direction gradually varies from 286° (February 15) to 269° (May 19). The
sunward feature in the northwest quadrant was morphologically very similar during
the entire time frame.
of m s-1 are not uncommon and were observed in other comets (e.g., comet Siding
Spring (C/2013 A1): Li et al. (2014), Tricarico et al. (2014), Kelley et al. (2014); comet
9P/Tempel 1: Meech et al. (2011), Vasundhara (2009)).
Considering that the corresponding geocentric distance during this interval is
approximately 4.0 – 4.3 AU and assuming a typical ground-based astronomical seeing
of 1 arcsec, the spatial resolution at the comet is ~3x103 km (for HST, the
corresponding spatial resolution is an order of magnitude smaller). During a time
interval corresponding to the single peak rotational period near 10.4 hours suggested by
Lamy et al. (2014), grains would have moved of the order of 4×102×cos
km in the
is the angle between the skyplane and the initial direction of the
feature. This distance is an order of magnitude smaller than the ground-based seeing
disk, so it is not surprising that we do not see any fine structure of the sunward feature
in the ground-based images. The angular resolution of the HST images is comparable
to the spatial displacement of grains in the sunward feature over a diurnal cycle.
Therefore, despite the better angular resolution, even for HST images, we do not have
sufficient spatial resolution to detect any diurnal scale variations in the coma structure.
On the other hand, the long-term variations over timescales of months in the sunward
feature during the mid-February to mid-May interval are generally consistent with the
evolution of the Earth-comet-Sun geometry and in particular the projected solar
direction with respect to the comet in the skyplane (Figures 1 and 5). Based on the
derived grain outflow velocity and the fact that we can observe the ultimate merging of
the sunward feature with the tail due to radiation pressure, we estimate that by the time
the grains reach the tail, they must have spent 2-4 weeks in the coma after being
ejected from the nucleus (cf. equation (7) of Mueller et al. 2013).
3.3 Dynamical Constraints Based on the Sunward Feature
In general, if a coma feature originates at or near a pole, the pole direction should lie in
a half-great circle defined by the PA of that feature. When the Earth direction as seen
from the comet changes due to orbital motions, the PA of the feature may also change.
One can use a range of PAs corresponding to observations made at different times to
determine a pole solution by considering the intersection of these half-great circles.
In the case of comet ISON, the Earth direction changed less than 8° from mid-February
to mid-May (cf. Figure 5). In addition, the lack of good spatial resolution of the sunward
feature due to the comparatively large geocentric distances to the comet resulted in
large error bars for the PAs for the sunward feature (cf. Figure 4). Therefore, from these
observations, the determination of a robust pole direction based on intersections of
multiple half-great circles corresponding to respective PAs is not feasible. In this case,
most half-great circles appear to intersect near the Earth’s direction, which are confined
close to each other. This result is simply a manifestation of the inadequacy of this
technique to yield reliable results when the Earth direction has not changed significantly
with time, and our data provide no additional constraints to the solution defined by Li et
al. (2013) in their Figure 4. This result was based on the higher resolution HST
observations (see Figure 5) with the assumption that the feature originated from a fixed
source region at or near the pole. As commented earlier in Section 3.1 (and also as
favored by Knight and Schleicher 2015), the feature could simply be the cumulative dust
outflow from the sunward side (with a significant contribution coming from regions near
the sub-solar point) and in that case the sunward feature cannot be used to constrain
the pole. Detailed modeling of this scenario is beyond the scope of this paper.
Figure 5. The family of pole solutions based on Li et al. (2013) is shown by the half-
great circle (with its width corresponding to the error in the determination of the PAs).
The Earth and Sun directions as seen from the comet from January 1, 2013 up to
perihelion are represented by blue solid and red dashed lines, respectively. The dots
denote the respective directions at 0 UT on January 1, April 1, July 1, October 1,
November 1, November 26, and November 28 while the direction of arrows denote
the change in directions with time. The dots for the Sun direction partially overlap
from January 1 through November 1 indicating that the Sun motion is minimal during
this time; the huge change in Sun direction from November 28 (0 UT) until perihelion
corresponds to <19 hrs. The green and dark gray parts of the half-great circle denote
the portions in sunlight and darkness respectively during the HST observations. The
Earth direction during the HST observations is nearly the same as that for April 1.
The rightmost end of the solid black line (which depicts the “most likely HST
solution”) represents the pole solution that is in the skyplane. The dashed grey line
represents the great circle solution from the November outburst observation
(Section 4.3) with the light grey envelope on the left indicating the range for the pole
solution for a 10° uncertainty in the PA and 30° uncertainty with respect to the
skyplane. For all the pole solutions in the figure, the diametrically opposite solutions
are also valid but are not shown.
4. Results during the Water-Dominated Phase
The coma campaign during the water-dominated phase focused on both continuum as
well as gas images. However, there were only a few observers who had access to
narrowband gas filters: Hoban and Prouty (Telescope at University of Maryland,
Baltimore County), Lister et al. (Faulkes Telescope North as part of the Las Cumbres
Observatory Telescope Network), and Knight and Schleicher (using various telescopes
at Lowell Observatory). The gas images of Knight and Schleicher were the only images
received by the campaign that had sufficient S/N to detect unambiguous coma features.
As these images are already discussed in detail in Knight and Schleicher (2015), we will
not consider them further. However, we point out that unambiguous gas coma features
were seen starting nearly a month prior to the perihelion (cf. Opitom et al. 2013b, Knight
and Schleicher 2015). In the remainder of this section, we will concentrate on the
behavior of the continuum features.
4.1 Searching for Coma Features in the Continuum
As September 2013 approached, the solar elongation increased sufficiently (to ~30°) for
ground-based observations despite the observing window on any given night being
short and the comet being at high airmass. Observations from September through
November were also facilitated by the decreasing heliocentric and geocentric distances
(Figure 1) and the resultant increase in the comet’s apparent brightness (e.g., Meech et
al. 2013). However, up until about two weeks prior to the perihelion (i.e., rh~0.65 AU),
we could not detect any unambiguous evidence of continuum features, except for the
dust tail, even with image enhancements. To illustrate that an unambiguous continuum
feature was not detected, we show in Figure 6, a high S/N image from October 17
(rh~1.3 AU) representative of the middle of the observing window when water was the
dominant volatile. This figure demonstrates the dependency of the “sunward feature” on
the assumed nucleus (center) location. As shown elsewhere using numerically
simulated images (Figure 9 of Samarasinha and Larson 2014), in the absence of any
corroborating evidence, one should be extremely cautious to trust the reality of an
apparent coma structure after image enhancement if that structure is sensitive to minute
changes in the chosen nucleus location such as a one-pixel offset. We carried out a
similar test for the September 12 images taken at the DCT by Knight and Schleicher
(2015) and in that case the feature is more stable even though the PA of the feature is
still sensitive to the chosen nucleus location. We attribute this behavior to the relatively
high S/N as well as the smaller pixel size of the DCT images. Even if the sunward
feature is real in early-September, we assert that its skyplane extent must be much
smaller than in the February to May time frame. I.e., at best, it can only be marginally
4.2 Why Were no Features Detected in the Continuum?
What happened to the sunward feature observed during the pre-water-dominated
phase? Did it disappear, as water became the dominant volatile responsible for
cometary outgassing? Since the change in the Sun direction as seen from the comet
from May to September is only about 3° (cf. Figure 5), it is unlikely that the lack of a
feature was caused due to a change in the insolation geometry.
If one assumes that (a) the sunward feature during the pre-water-dominated phase
retained the same character as during the water-dominated phase, (b) the change in the
dust outflow velocity V in the intervening period is small, and (c) the change in the
projection effects can be ignored6, then the sunward extent of the feature d should
decrease as rh
(cf. equation (4) of Mueller et al. 2013). In late-September when rh
6 For many of the “most likely solutions” suggested for the jet/pole direction in Figure 4 of Li et al. (2013)
(also Figure 5), the projection effects tend to make d even smaller during the September to October time
frame (compared with the February to May time frame) based on the corresponding Earth directions if the
feature is caused by a fixed source region on the nucleus.
Figure 6. A high S/N
continuum image of comet
ISON taken on October 17
(rh~1.3 AU) at the Liverpool
Telescope. The image is
enhanced using the division by
azimuthal average technique.
In the center panel, the
optocenter is taken to
represent the nucleus. In the
other panels, the center is
shifted by one pixel in the
respective directions. Each
panel is approximately 12,000
km across. The Sun is at a PA
of 113°. The dust tail is at a
direction. Note that the
apparent sunward feature in
the center panel cannot be
seen in all the panels at the
is nearly half of that in mid-May and
has nearly doubled, d should be ~700 km. As Δ ~
2.3 AU, the angular extent of the sunward feature should be ~0.4 arcsec which is
smaller than the typical astronomical seeing. Therefore, it is not surprising that we
cannot make a clear identification of the sunward feature in late-September even if it
indeed existed. In October and November,
increases further while rh decreases,
making the skyplane extent of the feature even smaller. HST images taken in the
continuum on October 9 (rh~1.5 AU) and November 1 (rh~1.0 AU) do not show a
sunward feature, or another feature for that matter, other than the dust tail (J.-Y. Li;
personal communications, 2013). This suggests that, at least around the times when
these HST images were taken, the skyplane extent of a sunward feature was extremely
small (<100 km). Another possibility is that such a feature was absent.
If the direction of the sunward feature was close to the skyplane during the mid-
February to mid-May time frame, then during late-September, the feature could still
maintain a PA approximately in the same general westerly to north-westerly direction as
that during the pre-water dominated phase. To illustrate this, in Figure 5, we show the
family of solutions suggested for the pole in Li et al. (2013) as well as the Sun and Earth
directions as a function of time. However, as seen from Figure 1, the PA of the solar
direction in the skyplane changed dramatically in July and by late-September the PA of
the Sun was around 110°. That means the February-May sunward feature would be
essentially in the tailward direction in late-September, thus effectively precluding its
detection. We emphasize that this argument is relevant only if the feature is due to a
fixed source region near the pole, the pole direction is close to the skyplane during the
mid-February to mid-May interval, and this direction was nearly the same in September.
4.3 Behavior of the Coma in the Month Before Perihelion
A series of observations indicating rapid increases in activity and outburst events were
reported in the month before perihelion. By November 5 (rh~0.9 AU), water production
rate increased by ~50% from two days earlier, and that of other gas species increased
by twice the corresponding amount, but no change was observed in the dust production
(Opitom et al. 2013b). On November 12 (rh~0.7 AU), water production rate was ~50%
higher than the previous night, and continued to increase by nearly an order of
magnitude over the next two days. Dust production also increased by an order of
magnitude by November 14 (Opitom et al. 2013a). Likely associated with this outburst,
two arclet-like “wings” were detected (Boehnhardt et al. 2013) in the coma for the first
time around November 14.2 UT, and confirmed on November 14.37 (C. Opitom;
personal communications, 2015) and November 14.99 (Ye et al. 2013). These “wings”,
which originate from the nucleus at PAs nearly perpendicular to the tail and curve back
in the tailward direction, are likely to be associated with fragmentation7 of the nucleus
(e.g., Harris et al. 1997, Tozzi et al. 1997, Farnham et al. 2001, Jehin et al. 2002,
Hadamcik and Levasseur-Regourd 2003, Boehnhardt 2004, Farnham 2009). A third
increase in the gas production was reported (Opitom et al. 2013c) to have started
7 Fragmentation here is referred to small pieces breaking away from the nucleus and not to the complete
disruption/breakup of the nucleus.
between November 18 and 19 (rh~0.5 AU). These observations reveal behavior that is
remarkably similar to that of comet D/1999 S4 (LINEAR) before its complete breakup.
One of the LINEAR outbursts was connected to a fragmentation event, with associated
“wings” in the coma, that may have been related to the final disintegration of the nucleus
(e.g., Farnham et al. 2001, Tozzi and Licandro 2002). The existence of “wings” in comet
ISON, after a long period of quiescence, may have foreshadowed the comet’s ultimate
Our global coma morphology campaign, which contains at least one set of images every
day from November 1-17, provides an opportunity to explore this outburst timeframe in
more detail. This task is challenging due to the non-uniformity of the data sets, but the
temporal coverage is of value. During and after the reported increase in gas production
on November 5, we see no changes in the dust coma morphology. However, the total
gas production in this event increased by only ~50%, and it was noted that the dust
production remained flat.
The November 12 outburst, on the other hand, showed an order of magnitude increase
in both gas and dust productions, indicating a significantly more energetic event that is
likely associated with the observed “wings”. Our campaign includes two sets of images,
obtained at the appropriate time and with sufficient S/N to investigate this timeframe.
Neither images acquired by Z.-Y. Lin and colleagues on November 13.85 nor images
from Q.-Z. Ye and colleagues on November 13.99 show any identifiable “wing”
structure, but images acquired by the same groups on November 14.86 and November
14.99, respectively, do show features similar to those reported by Boehnhardt et al.
(2013). Based on this, we suggest that the faint “wings” reported by Boehnhardt et al.
from images on November 14.2, first appeared in the coma sometime between
November 13.99 and November 14.2 UT. In Figure 7, we show a sequence of images
bracketing this time interval, showing the coma with no “wings”, the initial stage of their
formation during November 14, and well developed “wings” lasting for at least three
The "wing" structure, as noted above signifies that a fragmentation event has occurred,
and its timing corresponds to the (measured) “peak” in the production rates around
November 14 (Opitom et al. 2013a, Combi et al. 2014), rather than to the start of the
outburst on November 12. This indicates that the fragmentation is the result of the
increased gas production, rather than the cause of the outburst. The morphology of the
"wings" in both comet LINEAR and comet ISON, with nearly identical lobes in opposing
directions, suggests that the source region is near the equator of a rapidly rotating
nucleus, and the Earth is at low latitudes. These characteristics, combined with the
timing of the events, allow us to propose a possible mechanism to explain these
observations. As the nucleus approaches the Sun, heat penetrates into a pocket of
volatiles, dramatically increasing the gas production. The higher gas drag, aided by
erosion and centripetal acceleration near the equator, cause a piece of the surface to
break off, revealing fresh ices. Material from this new active area, rotating with the
nucleus, produces nearly axisymmetric lobes that appear as the "wings" seen in Figure
7, while contributing additional gas and dust to the peak of the outburst.
By November 15 and 16, the “wing” structure was mature and well established. The
dust in the “wings” further away from the nucleus was clearly affected by radiation
pressure, being pushed significantly towards the anti-sunward direction. Unfortunately,
we do not have sufficient observations to allow us to determine how long the “wings”
Figure 7. Images of comet ISON from November 12 (rh~0.70 AU) to November 16
(rh~0.56 AU) showing the development of the coma “wings” in comet ISON
subsequent to the outburst of November 12. The time of observation is indicated at
the top of each image and the time sequence is from left to right in the top row
followed by left to right in the bottom row. The nucleus is at the center of each panel.
Each panel is approximately 96,000 km across. The position angle of the Sun varies
from 112° to 110° during this interval. All images were enhanced using the division by
azimuthal median technique. The streaks present in some images are star trails. The
observers were: (a) N. Howes and colleagues, (b) Z.-Y. Lin and colleagues, (c) D.C.
Vogel, (d) C. Pruzenski and colleagues, (e) E. Gomez and colleagues, and (f) Z.-Y.
Lin and colleagues.
persisted past November 16. Nor can we evaluate the effects of the weaker outburst
reported by Opitom et al. (2013c) around November 18. They note the presence of
“wings”, though it is not known if these are newly generated in the outburst or if they are
simply the remaining vestiges of the November 14 features.
The equator being nearly edge-on around November 14 suggests that at that time the
pole solution should lie close to the great circle defined by the sunward and anti-
sunward directions as well as nearly in the skyplane, yielding a pole in the general
direction of (RA=285°, Dec=-20°). Figure 5 shows the great circle solution for the
sunward PA (dashed grey line) with a light grey envelope that denotes an uncertainty of
10° in the measured PA and an allowance for the pole being up to 30° out of the
skyplane. Although this solution overlaps the HST solution discussed earlier, they are
not required to be consistent because the pole could have changed due to the torques
caused by outgassing in the intervening period between April and mid-November (cf.
Samarasinha and Mueller 2013).
Many of our images from November 14 do not exhibit the “wings” seen in the high-
resolution data. This is likely due to a combination of differences in astronomical seeing
and S/N between images as well as the enhancement techniques being employed.
However, the overlap of these data represents a means of connecting the signature
from higher-resolution images to the lower-resolution images and for exploring the
effects of enhancing those images. The data taken at the TRAPPIST Telescope at the
La Silla Observatory in Chile around the same time as our observations show sharper
“wings” if all images are enhanced with the same technique (C. Opitom; personal
communications, 2015), which we attributed to the better seeing at the TRAPPIST
location. As for enhancements, Boehnhardt et al. (2013) used Laplace filtering which is
optimized for identifying spatial discontinuities. As a result, the strong sunward edge of
the coma seen for images from November 14 in Figure 7, where we have enhanced the
images by the more benign division of azimuthal median technique, may look like
“wings” if enhanced with a Laplace filter (e.g., see Figure 6 in Farnham 2009). (For a
detailed comparison of differences between enhancement techniques, the reader is
directed to Samarasinha and Larson (2014).) A comparison of different images
indicates that the same structure that shows well-developed “wings” in some
observations, appears as a flattening or squashing of the sunward side coma in others
(e.g., images acquired by D. Vogel on November 14.43 and C. Pruzenski and
colleagues on November 14.45). This type of behavior can be used as a signature of a
potential fragmentation event in future observations, when only low-resolution data are
An increase in polarization was observed for comet D/1999 S4 (LINEAR), which was
tied to its fragmentation (Hadamcik and Levasseur-Regourd 2003). However, to our
knowledge, no polarimetric observations are available from around mid-November until
ISON’s demise although studies were conducted by comparing coma colors (e.g., Li et
al. 2013) and polarization properties (e.g., Hines et al. 2014, Zubko et al. 2015) for
comet ISON during the first half of 2013.
5. Summary and Conclusions
We collected many hundreds of images from nearly two dozen groups of amateur and
professional observers, spanning nearly the entire time comet ISON was observable
from the ground. During November 2013, when ISON was brightest but its visibility was
severely restricted due to a small solar elongation, this allowed much better temporal
coverage than would have been possible from a single location. For most of the
apparition, comet ISON displayed no prominent morphological features. We attribute
this partially to the relatively large geocentric distances to the comet. The main results
based on this study are summarized below.
• A distinctive northwesterly sunward feature in the continuum marked the coma
morphology during the pre-water dominated phase. This feature did not vary on
diurnal timescales. We derive grain velocities of the order of 10 m s-1 and the
grains must have spent 2-4 weeks in the sunward side prior to merging with the
• During the water-dominated phase, either the earlier continuum feature was
absent or if it actually existed, it did not present itself prominently in the images.
I.e., its skyplane-projected extent was of the order of the astronomical seeing or
smaller and we provide an explanation for that based on the radiation pressure
effects and the ground-based observational capabilities.
• Nearly two weeks prior to the perihelion, the comet started to display a variety of
continuum features. The onset of this is related to the November 12 outburst
reported by other authors. We constrain the “wing”-like features to have
appeared between November 13.99 and 14.2 UT.
• This study shows that organized observing campaigns such as this one can
collect observations from numerous observers, both professional and amateur,
and assemble them into useful datasets. These campaigns may be most
valuable in situations where any single observer can only obtain data during a
small window of time, but contributions from many such observers provide
coverage that leads to a more complete understanding of the spatial and
temporal evolution of the comet.
The NASA Planetary Astronomy Program with grant NNX14AG73G supported NHS and
BEAM. MMK thanks NASA Planetary Astronomy grant NNX14AG81G. We thank the
CIOC and many other individuals for providing publicity about this campaign. We thank
C. Opitom for discussions related to their observations. MMK thanks Stephen Levine,
Brian Skiff, and Larry Wasserman for help obtaining the data from Lowell Observatory.
Z-YL thanks the staff of Lulin Observatory, Hsiang-Yao Hsiao, Chi-Sheng Lin, and
Hung-Chin Lin, for their assistance with observations. We thank the two anonymous
reviewers for their comments to improve the paper. This is PSI Contribution Number
Belton, M.J.S. 2013. The sources of the unusual dust jets seen in comet
103P/Hartley 2. Icarus 222, 653-661. DOI: 10.1016/j.icarus.2012.12.007.
Bockelée-Morvan, D., Crovisier, J., Mumma, M.J., Weaver, H.A. 2004. The composition
of cometary volatiles. In Comets II, (M.C. Festou, H.U. Keller, H.A. Weaver, eds.),
University of Arizona Press, Tucson, pp. 391-423.
Boehnhardt, H. 2004. Split comets. In: Festou, M.H., Keller, U., Weaver, H. (Eds.),
Comets II. University of Arizona Press, Tucson, pp. 301-316.
Boehnhardt, H., Tubiana, C., Oklay, N., Vincent, J.B., Hopp, U., Ries, C., Schmidt, M.,
Riffeser, A., and Gossel, C. 2013. Comet C/2012 S1 (ISON). CBET 3715.
Burns, J.A., Lamy, P.L., Soter, S. 1979. Radiation forces on small particles in the solar
system. Icarus 40, 1-48. DOI: 10.1016/0019-1035(79)90050-2.
Combi, M.R., Fougere, N., Makinen, J.T.T., Bertaux, J.-L., Quemerais, E., Ferron, S.
2014. Unusual water production activity of comet C/2012 S1 (ISON): Outbursts and
continuous fragmentation. Astroph. J. Letters 788, article id. L7, 5 pp. DOI:
Farnham, T.L. 2009. Coma morphology of Jupiter-family comets. Planetary and Space
Science 57, 1192-1217. DOI: 10.1016/j.pss.2009.02.008.
Farnham, T.L., Schleicher, D.G., Woodney, L.M., Birch, P.V., Eberhardy, C.L., Levy, L
2001. Imaging and photometry of comet C/1999 S4 (LINEAR) before perihelion and
after breakup. Science 292, 1348.
Farnham, T.L., Samarasinha, N.H., Mueller, B.E.A., Knight, M.M. 2007. Cyanogen jets
and the rotation state of comet Machholz (C/2004 Q2). Astron. J. 133, 2001-2007.
Farnham, T., Bodewits, D., Feaga, L.M., A'Hearn, M.F., Sunshine, J.M., Wellnitz, D.D.,
Klaasen, K.P., Himes, T.W. 2013. Imaging comets ISON and Garradd with the Deep
Impact flyby spacecraft. DPS 2013 #45, #407.08 (abstract).
Ferrín, I. 2014. The impending demise of comet C/2012 S1 (ISON). Planetary and
Space Science 96, 114-119. DOI: 10.1016/j.pss.2014.03.007.
Hadamcik, E., Levasseur-Regourd, A.-C. 2003. Dust coma of comet C/1999 S4
(LINEAR): Imaging polarimetry during nucleus disruption. Icarus 166, 188-194.
Harris, W.M., Combi, M.R., Honeycutt, R.K., Mueller, B.E.A. 1997. Evidence for
interacting gas flows and an extended volatile source distribution in the coma of comet
C/1996 B2 (Hyakutake). Science 277, 676-681.
Hines, D.C. and 11 colleagues 2014. Hubble Space Telescope pre-perihelion
ACS/WFC imaging polarimetry of comet ISON (C/2012 S1) at 3.81 AU. Astroph. J.
Letters 780, article id. L32, 6 pp. DOI: 10.1088/2041-8205/780/2/L32.
Howes, N., Guido, E. 2013. Comet Observations. Minor Planet Circular 81942.
Jehin, E. and 14 colleagues 2002. Split comet C/2001 A2 (LINEAR). Earth Moon
Planets 90, 147-151.
Kelley, M.S., Li, J., Mutchler, M.J., Weaver, H.A., Knight, M.M., HST ISON Imaging
Science Team 2014. The pre-perihelion size of the nucleus of comet C/2012 S1 (ISON).
AAS Meeting #223, #247.15 (abstract).
Knight, M.M., Battams, K. 2014. Preliminary analysis of SOHO/STEREO observations
of sungrazing comet ISON (C/2012 S1) around perihelion. Astroph. J. Letters 782,
article id. L37, 5 pp. DOI: 10.1088/2041-8205/782/2/L37.
Knight, M.M., Schleicher, D.G. 2015. Observations of comet ISON (C/2012 S1) from
Lowell Observatory. Astron. J. 149, article id. 19, 15 pp.
Knight, M.M., Walsh, K.J. 2013. Will comet ISON (C/2012 S1) survive perihelion?
Astroph. J. Letters 776, article id. L5, 5 pp. DOI: 10.1088/2041-8205/776/1/L5.
Knight, M.M., Schleicher, D.G., Farnham, T.L., Schwieterman, E.W., Christensen, S.R.
2012. A Quarter-century of observations of comet 10P/Tempel 2 at Lowell Observatory:
Continued spin-down, coma morphology, production rates, and numerical modeling.
Astron. J. 144, article id. 153, 25 pp. DOI: 10.1088/0004-6256/144/5/153.
Lamy, P L., Toth, I., Weaver, H.A. 2014. Hubble Space Telescope observations of the
nucleus of comet C/2012 S1 (ISON). Astroph. J. Letters 794, article id. L9, 6 pp.
Li, J.-Y., Kelley, M.S., Knight, M.M., Farnham, T.L., Weaver, H.A., A'Hearn, M.F.,
Mutchler, M.J., Kolokolova, L., Lamy, P., Toth, I. 2013. Characterizing the dust coma of
comet C/2012 S1 (ISON) at 4.15 AU from the sun. Astroph. J. Letters 779, article id. L3,
5 pp. DOI: 10.1088/2041-8205/779/1/L3.
Li, J.-Y., Samarasinha, N.H., Kelley, M.S.P., Farnham, T.L., A'Hearn, M.F., Mutchler,
M.J., Lisse, C.M., Delamere, W.A. 2014. Constraining the dust coma properties of
comet C/Siding Spring (2013 A1) at large heliocentric distances. Astroph. J. Letters 797,
article id. L8, 7 pp. DOI: 0.1088/2041-8205/797/1/L8.
Lisse, C.M., and 13 co-authors 2013. Observing comet C/2012 S1 (ISON) with Spitzer.
DPS 2013 #45, #407.06 (abstract).
Meech, K.J., and 57 co-authors 2011. Deep Impact, Stardust-NExT and the behavior of
comet 9P/Tempel 1 from 1997 to 2010. Icarus 213, 323-344. DOI:
Meech, K.J., and 28 co-authors 2013. Outgassing behavior of C/2012 S1 (ISON) from
2011 September to 2013 June. Astroph. J. Letters 776, article id. L20, 6 pp.
Mueller, B.E.A., Samarasinha, N.H., Farnham, T.L., A'Hearn, M.F. 2013. Analysis of the
sunward continuum features of comet 103P/Hartley 2 from ground-based images.
Icarus 222, 799-807. DOI: 10.1016/j.icarus.2012.08.014.
Novski, V., Novichonok, A. 2012. Comet C/2012 S1 (ISON). CBET 3238.
Opitom, C., Jehin, E., Manfroid, J., and Gillon, M. 2013a. Comet C/2012 S1 (ISON).
Opitom, C., Jehin, E., Manfroid, J., and Gillon, M. 2013b. Comet C/2012 S1 (ISON).
Opitom, C., Jehin, E., Manfroid, J., and Gillon, M. 2013c. Comet C/2012 S1 (ISON).
Samarasinha, N.H., Larson, S.M. 2014. Image enhancement techniques for quantitative
investigations of morphological features in cometary comae: A comparative study.
Icarus 239,168-185. DOI: 10.1016/j.icarus.2014.05.028.
Samarasinha, N.H., Mueller, B.E.A. 2013. Relating changes in cometary rotation to
activity: Current status and applications to comet C/2012 S1 (ISON). Astroph. J. Letters
775, article id. L10, 6 pp. DOI: 10.1088/2041-8205/775/1/L10.
Schleicher, D.G., Woodney, L.M., Millis, R.L. 2003. Comet 19P/Borrelly at multiple
apparitions: Seasonal variations in gas production and dust morphology. Icarus 162,
Sekanina, Z., Kracht, R. 2014. Disintegration of Comet C/2012 S1 (ISON) shortly before
perihelion: Evidence from independent data sets. arXiv:1404.5968.
Tozzi, G.-P., Cimatti, A., di Serego Alighieri, S., Cellino, A. 1997. Imaging polarimetry of
comet C/1996 B2 (Hyakutake) at the perigee. Planetary and Space Science, 45, 535-
Tozzi, G.P., Licandro, J. 2002. Visible and infrared images of C/1999 S4 (LINEAR)
during the disruption of its nucleus. Icarus, 157, 187-192.
Tricarico, P., Samarasinha, N.H., Sykes, M.V., Li, J-Y., Farnham, T.L., Kelley, M.S.P.,
Farnocchia, D., Stevenson, R., Bauer, J.M., Lock, R.E. 2014. Delivery of dust grains
from comet C/2013 A1 (Siding Spring) to Mars. Astroph. J. Letters 787, article id. L35, 5
pp. DOI: 10.1088/2041-8205/787/2/L35.
Vasundhara, R. 2009. Investigations of the pre-Deep Impact morphology of the dust
coma of Comet Tempel 1. Icarus 204, 194-208. DOI: 10.1016/j.icarus.2009.06.001.
Ye, Q.-Z., Hui, M-T., Gao, X. 2013. Comet C/2012 S1 (ISON). CBET 3718.
Zubko, E. and 12 colleagues 2015. Comet C/2012 S1 (ISON) coma composition at ~4
AU from HST observations. Planetary and Space Science, in press.