Imhotep-Erosion at accumulation basin B


Yellow heart shape- the focus that instigated the erosion in accumulation basin B. 


Yellow- the post-perihelion scarp that’s remained behind in the location of the focus and is more obvious after the drop in levels in front of it due to erosion. 

The two photos above are reproduced below with their full keys and unannotated versions under the sub-heading ‘The Photos’. As such they are counted here as being the header and are not numbered.


This is the third page in a series of three pages. They’re called pages by WordPress when they’re placed in the menu bar and are not part of the run of chronological blog posts. However, they should be regarded as medium-length blog posts and will be upgraded to blog post status in due course. Right now, they would interfere with the theme prevailing in the current chronological order of the main posts. The first two pages of the three are here:

“Focus for Groussin et al. Area B found”

“Ice Signature at the Area B Focus”

All three pages are in response to the Rosetta blog’s request for contributions regarding any changes on the comet during the perihelion period in 2015. That Rosetta blog post is here:

The perihelion date was August 13th 2015. The pictures above and below were all taken well before or well after that date. 

The two scientific papers that are referred to in this post and the other two pages are Groussin et al. 2015 and Auger et al. 2015. They’re linked under the references section at the bottom and are free to download. 

The three pages are linked in terms of theme and build-up of evidence. It would be best to read the other two first, but the most important themes from those pages are imported here. Key terms like ‘focus’ ‘area B’ and ‘ice signature’ refer to those features as defined in the two preceding pages. They are defined again as we go along but not necessarily at their first mention. 

This page is concerned with a moderately large area of erosion in accumulation basin B on Imhotep during the perihelion of 2015. It wasn’t mentioned in Groussin et al. 2015 or anywhere else as far as I know. Accumulation basin B is the name given by Auger et al. 2015 for the middle, oval-shaped basin in the west of Imhotep. 

Note that we shall be talking about two areas that have been officially named in scientific papers and both have a ‘B’ in them. It’s worth heading off any future confusion at this point so when you see ‘area B’ it’s the area of erosion out on the smooth terrain that is next to but not part of the accumulation basin. It was referred to at length in the first two pages. When you see ‘accumulation basin B’ or ‘basin B’, it refers to the actual basin of course. Basin B wasn’t mentioned in the other two pages. Without this clarification, one could easily read ‘basin B’ and think it means ‘area B’, especially seeing as area B is, topologically, a basin. Basin B is often just called ‘the basin’.

It had been intended that the first two pages were going to provide evidence for the erosion processes in this page. It was a hypothesis mooted in the conclusion of the second page: that the accumulation basin B erosion spread westwards from the same apparent ice signature that the area B focus had started out from. But since publishing page two, the actual focus for the accumulation basin erosion has been found anyway. It’s in the Groussin et al 2015 photos and I simply hadn’t noticed it before. They show that it’s at the opposite end of the basin and that the erosion spread eastwards from there and not westwards. The basin B focus appeared for the first time in the 27th June 2015 photo.

This impromptu discovery of the basin’s erosion focus has deflected the build-up of evidence in a rather different direction. The eastward erosion that’s documented in the Groussin et al. photos will be described without any need to invoke evidence from pages one and two. But, since the eastward progression appears to have petered out and not quite reached the B focus ice signature, the original hypothesis is invoked after all for that small, remaining area. So pages one and two were needed after all but don’t take centre stage. 

As mentioned in the other pages, the term ‘focus’ is coined here but was not used in Groussin et al. 2015. It’s defined as being the small starting area where the erosion for each larger area of erosion (and consequent subsidence) was instigated. Groussin et al. 2015 established the exact foci for four of the five areas of erosion that they documented (areas A, C, D and E). Page one of this series identified the exact focus for area B. 

The newly discovered area of erosion in basin B is a sixth area so it’s in addition to the other five areas, A to E, out on the smooth terrain. The focus in basin B is a stand-alone feature, clearly depicted in the photos, without any need of corroborating evidence. Similarly, its growth, that is, the eastward erosion front movement across basin B, is also depicted in the Groussin et al photos, so it doesn’t need any corroborating evidence either. 

However, the eastward erosion event in the basin is nevertheless intimately related to the features and processes in the first two pages. It’s not physically related to those features but related in terms of the processes at play being similar: starting from an east-facing scarp and sweeping eastwards. Also the fact that the erosion front was curved and the depth of subsidence was similar to the 5 metres or so depths for areas A to E, possibly a little deeper. 

The discovery is elaborated on under the ‘An Impromptu Discovery’ subheading below. 


Photo 1- a pre-perihelion NAVCAM photo showing a curve in yellow dots. This is called the pre-perihelion curve in this post and is mentioned a lot so it’s probably worthwhile trying to memorise it as well as the three faint curves between it and the cliff behind it (not annotated). That section of cliff incorporates the B focus ice signature. Together, this stack of curves appear to run the length the B focus ice signature. 


Photo 2- Post-perihelion. This photo was taken on May 11th 2016 by the OSIRIS narrow angle camera so it’s nine months after perihelion. You can see that the pre-perihelion curve in photo 1 has disappeared but now a new, curved scarp has appeared beyond it. This is called the post-perihelion curve in this post and again, it’s probably worth memorising it or at least remembering that photos 1 and 2 show the two important curves. 


Photo 3

This is photo 1 reproduced with both curves annotated to show we’re looking at two distinct curves, one pre-perihelion (and actually there in this photo) and another post-perihelion (not there because it appeared later). However, if you study the unannotated versions below, you can see just the faintest of shadow prefiguring the later curve (the lower-left one). That curve clearly became deeper, steeper-edged and so more pronounced. The three faint, concentric curves can again be seen behind the pre-perihelion curve but aren’t annotated. 

Photo 4

This is photo 2 reproduced with both curves as well. In this case, the real curve is the post-perihelion one, further from us, and the now-disappeared, pre-perihelion curve is the one nearer to us. However, there appears to be the faintest hint of its former existence. It’s annotated anyway so as to show its former position, pre-perihelion when it was more defined. Similarly, the three concentric curve position apexes are are marked where they most likely were. Again, there may be the faintest hints of their former lines so those have been chosen as the apexes but it’s not at all certain that these subtle features really are the vestiges. Either way, the apex dots are within ten or fifteen metres of the actual lines in the pre-perihelion version. One can tell that from the fiduciary points along the rocky perimeter and the boulders.

Photos 5 and 6- the unannotated versions.

Photo 7- This is a pre-perihelion photo from Auger et al. 2015, taken in 2014, long before perihelion. It’s for location context only. 


The orange letters denote what are referred to as accumulation basins in Auger et al. 2015. The erosion described in this post is in basin B. The post-perihelion curved scarp sits right under the ‘B’ but is not actually there in this picture because it’s a pre-perihelion photo. The pre-perihelion curve is next to the rocky perimeter along a straight line between B and D. It’s actually there but not very obvious in this picture. 

Photo 8- the accumulation basin B focus appears (27th June 2015). Photos 8, 9 and 10 have their unannotated versions below them. 


Red- cliff perimeter bordering the smooth terrain. 

Yellow heart shape- the accumulation basin B Focus. 

Large yellow area at top- area B in Groussin et al. 2015 (growing as well). 

Light blue- the ice signature at the area B focus.

Small yellow dots above the ice signature- the area B focus. 

Small yellow curve below the ice signature- the pre-perihelion curve position. 

Photo 9- the eastward progression of the accumulation basin B erosion front (2nd July 2015). 


Dark blue- these are essentially the red perimeter but they run along the tops of cliffs and hills that are obscuring the true cliff base behind them. That’s because this photo has a low incidence angle. The blue-dotted parts add a 3D component to the depicted perimeter. 

Light blue at top left- the two isolated boulders out on the smooth terrain. These were dotted blue in previous pages because they were next to the area E focus. They are dotted here simply for orientation. 

Other colours- these are same as for photo 9. However, note that the heart shape has become egg-shaped with the post-perihelion curved scarp (smaller curve) distinct from the eastward-moving, curved erosion front. 

Photo 10- the further eastward progression of the basin B erosion front (11th July 2015). 


This shows the eastward front changing from a curve to a point as if it was running out of steam. The point kisses the pre-perihelion curve at the boulder.  

Photo 11- this is photo 8 reproduced, showing the post-perihelion curve superimposed on the heart. The erosion worked its way round the left hand section of the curve, as shown in photo 9, and then onwards and eastwards. The heart extends beyond the curve on the right, suggesting the erosion went southwards a little way too. However, the post perihelion photo (photo 2) doesn’t seem to show this southern side as an obvious scarp like it seems to be here. Photo 2 shows only the curve as superimposed here. 

Photo 12- this is photo 10 reproduced, showing the post-perihelion curve in relation to the pointed, eastward erosion front. 


This sub-heading repeats some of the things mentioned in the introduction. Sometimes that’s not a bad thing. Although this isn’t difficult to understand per se, it is convoluted in terms of the different shapes, times and erosion directions, photos etc.. One badly worded sentence can send you halfway across the comet, scratching your head as to how you got there. 

This page had to be rewritten, owing to the new find in the Groussin et al. 2015 photos. This find is the previously missed focus at the back of basin B. It was found after publishing page two of the series of three. In the conclusion of that second page, there was a reference made to the possibility that the ice signature at the area B focus also instigated all the erosion in accumulation basin B. That erosion would be moving across the accumulation basin in the opposite direction to area B’s erosion front vector across the larger area of smooth terrain. The erosion front vector is defined here as the mean direction of travel of the scarp as it eats its way across the basin or smooth terrain. 

However, the new discovery in the Groussin et al. 2015 photos clearly shows the erosion starting from a small heart shape at the back of the basin, which would be its western end. It seemed to start at or very near a shallow, curved ridge or scarp. That is the post-perihelion curve for this post. The first photo to record it is the one dated 27th June 2015. Whether it was noticeable as a scarp before the erosion started is debatable. What’s for certain is that it is now a much more noticeable scarp due to the erosion it instigated. The erosion progressed predominantly eastward from the scarp but it didn’t all go one way. There was a small southward component that seems to have moved in that direction from the top of the scarp. It’s assumed here that the scarp instigated the erosion but whether it did or not, the focus, the erosion and the scarp really did appear as evidenced in the photos. So speculation on the scarp instigating the erosion, i.e. being the exact focus, is secondary to the fact that all three really occurred. 

Photos 8, 9 and 10 were taken from the originals used in the Groussin et al. 2015 paper figure 1 montage. Those originals were posted in a Rosetta blog post reporting on the Groussin et al. paper. That blog post is here:

This discovery of a new focus and erosion front vector is, on balance, a step forward because it supersedes the mere hypothesis that the area B focus’ ice signature might have been responsible for the accumulation basin erosion. It replaces that hypothesis with the certainty that the heart-shaped focus at the back of the basin was responsible. The erosion front moved eastwards across the accumulation basin which is in the opposite direction to the hypothesised, focus-B-induced front.

This newly confirmed erosion front vector is in keeping with the erosion front vectors for areas A and B in Groussin et al. 2015. Those two areas also had foci (start points) at their western ends and those small initial areas grew essentially in an eastward direction whilst becoming somewhat wider to the north and south in the process. So the newly discovered focus and erosion front vector for accumulation basin B behave in a very similar way to the two main areas, A and B, out on the smooth terrain. This corroborates the the finding in Groussin et al. that erosion is instigated from east-facing scarps and cliffs and then gobbles its way in a preferred direction, which is essentially eastwards.


With regard to the first two pages in this series of three, they were admittedly working up to the proposition that the B focus instigated the erosion across the basin and that theory has now received a severe blow. 

However, there are three positive factors that emerge. Firstly, page one found the focus for area B, which is itself a stand-alone find, hopefully of some worth. 

Secondly, page two found a preponderance of ice at that focus, suggesting that for area B at least, it was the ice signature that kicked it off. So that was an interesting but not conclusive find. If correct, it will corroborate the ice signatures found by Groussin et al. 2015 at the foci for areas A, C, D and E.

Thirdly, the photos above show evidence of erosion in the basin. That evidence had initially been couched in long arguments and counter-arguments concerning the possibility that it was just lighting/shadow effects. That discourse has now been edited out because the newly discovered focus (and the two additional Groussin et al. photos depicting its growth) means that the erosion we see in the all the photos in this post is unequivocal. This means we have not only the ‘before’ and ‘after’ photos (photos 1-6 above) but a clear growth vector and chronology as well (photos 8-10).

Moreover, although the theory that the area B focus instigated the basin erosion has been dealt a blow, it’s not necessarily dead. This is because the heart-shaped focus grew eastwards but then appeared to be petering out in the last photo (photo 10) as it attempted to reach the B focus at the other end of the basin. Yet, in the post perihelion photo (photos 2 and 4) there does seem to be erosion all the way across, right up to the cliff edge above the B focus. The point at which the eastward, heart-shaped erosion appeared to peter out was exactly where it kissed the pre-perihelion curve apex. There’s a boulder where they kiss. This suggests that whatever erosion mechanism caused the pre-perihelion curve in the first place, it was sweeping over that area again in perihelion 2015. It swept that small area between the cliff and the pre-perihelion curve and did so in a manner that made it meet up and merge with the eastward-eroding component. That’s why the pre-perihelion curve is now gone. 

So it would still be appropriate to argue for the case that the erosion front was instigated at the B focus ice signature and moved that short distance westward to merge with the eastward erosion front vector across the basin. It could equally be argued that the eastward erosion front picked up steam again (after its last documented position in photo 10) and swept all the way to the cliff top above the B focus. 

If the B focus ice signature is indeed responsible for that short westward erosion front, it would be moving away from the side of the ice signature that’s opposite to the area B focus. So it would have started from the top of the small cliff. The area B focus itself was directly below, at the bottom of the cliff where it bordered the smooth terrain. The actual cliff face area in question is about 50 metres long (that is, the area B focus length consisting of two 25-metre-diameter circles at the time) and roughy between 20 and 30 metres high.

Whichever scenario played out, eastward or westward erosion, the before and after photos constrain this small area of erosion between the cliff top and the pre-perihelion curve as being during perihelion 2015, just like the eastward erosion from the initial heart-shaped focus.

Photo 13 shows the case for the westward erosion from the focus B ice signature to the pre-perihelion curve. 

Photo 13, showing the Groussin et al. area B focus ice signature sitting right behind the main pre-perihelion curve and its three nested companion curves that stack back to the focus. 



Small yellow- this is the perimeter of the area B focus as it would later appear on June 5th 2015 just before perihelion. So in this much earlier photo, it’s just the future outline of the yet-to-appear focus. 

Large yellow- the pre-perihelion curve with its three nested curves (not dotted) between it and the ice signature. 

Light blue- the area B focus ice signature. These are the three main signatures as depicted in the second page of the series. 

The small size of the area B focus (50 metres long along the perimeter and 20-25 metres wide metres out from it) shows that it’s enclosed within the length of the pre-perihelion curve and its three companion curves that stack back to the cliff top. 

This suggests the main pre-perihelion curve had already eroded the short distance from the focus on the perimeter. It had eroded across to where it is found in the above photo which is dated well before perihelion 2015. This probably happened during previous perihelions as no mention was made in Groussin et al. 2015 of this erosion front moving from the focus to its position in the photo during 2014 or early 2015. I also didn’t notice any such change during that period. Further evidence for the pre-perihelion curve eroding all the way from Groussin’s area B focus is the fact that its end is clearly touching the rocky perimeter and the right hand extremity of the focus (the other end is in shadow). Also, its three concentric curve companions suggest the presence of previous erosion fronts. 

So the main curve might be considered to have eroded from the rocky perimeter in a mirror image to its larger twin (area B) that eroded its way out across the larger smooth area and in the opposite direction. If this is the case, it eroded as far as the pre-perihelion curve where it met up with the known eastward erosion front. Otherwise, the eastward erosion carried on all the way to the cliff top. We know this area eroded in one direction or the other because photo 2 shows that the pre-perihelion curve has disappeared. 


Photo 14- The 2° phase angle photo. This was taken in April 2016, long after perihelion in August 2015. The post-perihelion curve is dotted yellow.


This photo has a 2° phase angle so it’s almost zero phase. Zero phase means the sun is behind the camera and so no shadows are present directly under the camera. 

The post perihelion curve is visible here. It’s also called ‘the scarp’ in this sub-heading, both for brevity and to emphasise its appearing due to the erosion around it. It’s also the curve along which the heart-shaped focus progressed. In other words, it was initially noticed simply as a new scarp with significant depth, betraying new erosion. Then the true focus for the basin B erosion was found in the Groussin et al. 2015 photos. It was only after the focus discovery that I realised the focus and the post perihelion curve were one and the same thing or at least very close to each other and overlapping. 

The post-perihelion curve is unusually white in this photo. In the larger, 0°, zero phase photo (not shown), all the classic Imhotep ice signatures were also white. Since all the classic ice signatures (established via 481nm wavelength colour ratio analysis) are unusually white in zero phase, it suggests that the post-perihelion curve is icy. Since the curve is also the basin B focus, it corroborates Groussin et al’s suggestion that erosion for areas A to E started from icy scarps or cliffs. It’s also east-facing as they noted for the cliffs and scarps for areas A to E. 

Incidentally neither the Auger et al. 2015 nor the Groussin et al. 2015 photos can tell us about any potential ice signature at the accumulation basin B focus. Auger et al’s blue ratio picture was taken long before perihelion and we know that it depicts the foci for areas A, C, D and E. They show no ice signature at that time. This is despite those areas having obvious ice signatures at their foci during perihelion (per the Groussin et al. blue ratio photos). The key to this paradox is, as Groussin et al. suggest, that a covering of dust was disguising the signatures. So the accumulation basin B focus signature (the curved scarp) could equally have been covered with dust when the Auger et al. blue ratio photo was taken. As for the Groussin et al. blue ratio photos, they don’t include the Basin B focus (or perhaps they may have done so but were cropped). Only their ordinary context photos show the focus and its eastward progression. That’s why the paragraph above uses the behaviour of the known, classic ice signatures as a proxy for a rough calibration of the ice signature strength along the basin B scarp. 

So, if it is an icy scarp, it explains why the post-perihelion curve is so obvious in the 2° phase angle photo. The erosion worked its way round the curve, as we know, but the scarp itself was more resistant and rock-like than the area in front of it that experienced runaway erosion. That erosion took off once the scarp had kicked it off, just like the scenario for the instigation of areas A to E out on the smooth terrain. The consequent subsidence left the curved scarp ‘high and dry’. This hypothesis, which is really a restatement of Groussin et al’s hypothesis for instigation, is loosely related to the process in the next sub-heading. 


Photos 15 and 16.



Photos 15 and 16 unannotated.

The single fuchsia dot in both photos denotes a rock or tiny outcrop which appeared after perihelion and which is located between the pre- and post-perihelion curves. Its position is marked on the pre-perihelion version despite not being visible as a rock. This is because, on closer scrutiny of the unannotated version of that photo, it shows the faintest of curves where this new rock appears later. In fact, there is the faintest of double bumps next to the pink dot in the pre-perihelion version so perhaps this is the rock, sitting under the dust. But it certainly looks a lot more obvious in the post-perihelion version. That fact in addition to the known erosion front that swept over it suggests that it was substantially more exposed via the perihelion erosion event. 

The alignment of the rock in both photos was established by the fact that it sits on a line that runs between the large boulder at the northern end of the post perihelion curve and the apex of the 90° turn in the perimeter cliff. 

The new rock or outcrop feature also appears to be curved around its far (western) side as viewed in the photo. It looks as though it nestles into that faint pre-perihelion curve location. Again, it’s better to look at the unannotated versions after familiarising yourself with the details in the dotted versions. 


Photos 17, 18 and 20 show the curved scarps in this post from a different viewpoint. It’s a close-up from the Auger et al. 2015 figure 7 photo. Photo 19, shows the curves plus the Auger et al. ice signatures copied over.

Photo 17


Photo 17 shows the unannotated zoom showing the relevant area. You can see mauve ice signatures dotted here and there.

Photo 18

Photo 18 shows the two curves, the area B focus and also highlights the main ice signatures in light blue. Larger rocks or stronger signatures have bigger dots but it’s a somewhat subjective interpretation. These ice signatures may eventually guide us as to why the focus started where it did. The area B focus ice signatures are the three at bottom right including the main one which is the larger dot, cut in half (but see clarification below). 

Photo 19

Photo 19 is the same as photo 4 with the two yellow curves but with exactly the same ice signatures dotted blue as well. 

Photo 20

The red line in photo 20 is the footprint of the view in photo 19. On close scrutiny, you can see that the large dot that was cut in half (half off-frame) in photo 19 was actually shunted about one dot width to the left in order to peep into view and betray this important ice signature’s position. It really should be off-frame in photo 19. It’s the main signature above the area B focus so it seemed sensible to shunt it over and then clarify its true position just off-frame in photo 20. You can also see that there are two small blue dots below the main signature, down on the cliff base. There was only one dot here in photo 19 because the other one was off-frame. This was also annotated as one dot in the second page of the series because they are so close together. It’s worth checking the original because, as mentioned above, it’s rather subjective. The signatures seem look stronger as you zoom out and are somewhat elusive if casting around at full zoom. 

Finally, the three tiny dots just above the half-dot in photo 19 should really be just off-frame too.



There are two items of evidence laid out above that show, in retrospect, that the post-perihelion scarp was destined to be the future site of erosion. These are:

a) the zero-phase photo showing the presumed ice signature matching up nicely along the line of the post-perihelion scarp. The zero-phase photo was taken in 2016, after perihelion so it was showing the actual eroded scarp and the now-revealed ice signature. 

b) the second item of evidence is the fuchsia rock that betrayed a sympathetically concentric curve to the main scarp (although, true to say, it’s much smaller but nevertheless orientated in the same direction). This would be flimsy evidence indeed if it were presented, pre-perihelion, as evidence for a prediction of a scarp forming further back in a few months’ time. However, if we look at the original pre-perihelion photo very carefully, we can see a concentric curve pattern on the surface between the small fuchsia rock curve and the new scarp curve.

Photo 21- the concentric curve pattern lying between the fuchsia rock curve and the post-perihelion, curved scarp (not shown). The fuchsia dot is on the fuchsia rock so the blue curve it’s sitting on is the fuchsia rock curve. 

Photo 22- this is the same as the photo above but it’s the pure pattern without the fuchsia dot. You’re strongly advised to check the original below for yourself. The curves are made up of small polygons and although the overall pattern is clearly concentric curves, some of the more irregular polygons can kick you between neighbouring curves. So it’s a judgment call teasing the curves apart and which larger, more irregular polygons belong to which obvious curve either side of them. A case in point is the photo below this one that shows the most obvious curve of all across the main erosion front. The slightly less obvious right hand tip has jumped to the end of the next curve when compared with the photo above. 

Photo 23- The most obvious curve with the pre- and post-perihelion curves added. It seems to be matching up well to the future scarp beyond it. 


The concentric curve pattern is made up of concentric chains of very small polygons. Many are quasi square, some rectangular, and just a few are more non-descript in terms of a formal shape identification other than being irregular polygons. The irregular ones are nevertheless obviously there, trying to merge into the concentric pattern but messing it up a bit in the process. All shapes have rounded corners, even the squarest of the squares. The curves that are formed by the polygons are stacked fairly tightly together across accumulation basin B between the fuchsia rock curve and the main post-perihelion curve. The concentric chains are not only fairly concentric within their own pattern but also concentric with the fuchsia rock curve behind them and the post-perihelion curve ahead of them. 

The concentric pattern fattens more and more as we progress from the tight fuchsia rock curve to the fatter post-perihelion curve. So even though the fuchsia rock curve is much tighter, this is how the ‘concentric’ pattern matches both a fat curve and a tight curve. 

These curved chains of polygons therefore appear to be the precursors of the soon-to-be scarp. The last two curves in the pattern, towards the back of the basin, certainly have a similar shape to the future scarp that’s located near the curve of rocks. That location is just beyond the last polygon curve. The penultimate polygon curve is the most obvious one that’s shown in photo 23. The last one is the slightly tighter one with a central bull nose that appears to be nosing in towards the curve of rocks at the back but not quite reaching the post-perihelion scarp location. 

It’s highly significant that this last curve with the bull nose is that shape and in that location. It’s the exact location and shape of the back edge of the ice signature in the zero-phase photo above. It follows the back perimeter of the signature thus:

Photo 24 plus original- the bull-nose curve follows the back perimeter of the ice signature.

Note that the blue bull-nose curve, which is the last polygon curve, follows the back of the ice signature. The ice signature is the more obviously white area and the reference to the “back” location is with respect to the “forward” direction of the curves. The yellow post-perihelion scarp follows the front of the ice signature and is also bull-nosed. This annotation of the post-perihelion scarp is far more accurate than any of those shown above. Those earlier photos show the yellow curve skirting round just in front of the actual bull-nosed curve that forms the post-perihelion scarp. They just depict the average curve. This photo follows the actual bull nose of the post perihelion scarp because we’re drawing attention to how it’s so similar in shape to the last blue polygon curve. The two bull-noses match in a translational sense (they nest perfectly). Furthermore, the two lines book-end the ice signature. It could hardly be neater and it shows that the pre-perihelion pattern of curves certainly appear to prefigure the future erosion across basin B. The only thing it couldn’t prefigure is that the erosion would overshoot the pattern perimeter (the blue bull nose) by a few tens of metres whilst maintaining the same-shaped erosion front. 

We should also bear in mind that the more accurate annotation of the yellow post-perihelion scarp in the above photo should supersede all the less accurate ones above. This is because it’s an overhead view so the bull nose in the scarp is more accurately portrayed. The close-up (ordinary OSIRIS) photo at the top of this post is at a low angle of incidence. This is good for seeing the scarp, almost face on, but not so good for discerning the exact perimeter shape other than seeing that it’s a curve. That said, the bull nose is nevertheless visible anyway in that photo which is why we can have confidence in flying up to the vertical view in the zero-phase photo and using that for better guidance of the perimeter shape. We’re therefore using the two photos together: the first to prove there is indeed a scarp and the second to delineate the exact perimeter of that scarp. The second can’t prove there’s a scarp because it’s an overhead view. 

So to recap, the concentric polygon pattern stops where the ice signature starts. The pattern gives way to the ice along a very defined line as can be seen from photo 24. That line is the last concentric line in the pattern, the blue bull-nose line. In other words the ice signature was perhaps an intact strip of still unaffected ice whereas the polygons in the pattern betray areas that may have been partially affected already by sublimating gases coming through the dusty surface. This would also explain why there’s less or no ice signature on the polygon pattern area in the zero-phase photo. It would have been partially used up in previous perihelions, leading to full collapse and elimination of ice in 2015. That would have exposed the as-yet unaffected bed of ice due to the bull-nose curve collapsing in front of it. It would be a new, east-facing scarp that then eroded that small distance from the bull-nose curve to the post-perihelion scarp just beyond it, next to the curve of rocks. As mentioned above, the post-perihelion scarp delineates the front perimeter of the ice signature which is also bull-nosed. This suggests that the erosion progressed along a front that was largely unaltered in terms of shape as it eroded across the ice signature. 

So the erosion progressed through the polygon pattern and also through the the newly exposed ice strip beyond the polygon pattern. However being newly exposed, the ice perhaps didn’t have time to be exhausted during 2015, hence its signature in the zero-phase photo.


These photos are included so as to nail down the read-across from the pre-perihelion photo to the post-perihelion ice signature photo. This is to make sure we got the curves in the right place. 

The photos use several boulders as fiduciary points. They appear in both photos. These boulders are marked, unusually, in red. Boulders are usually marked light blue, because they’re often icy, but the polygon pattern was itself marked light blue because it too is suggested to betray past sublimation from ices below. So the red for the fiduciary boulders is an expedient for this post only so as to avoid confusion. 

Photos 25/26- close up of the pre-perihelion and post-perihelion photos. They’re together for toggling between. 

In photos 25/26, the blue curves are shown in the pre photo and the post-perihelion scarp is shown in the post photo. Note that it’s the old, rough yellow curve with the true bull nose of the new scarp just inside it. The red boulders are the same in both photos. The small perimeter of red dots is a tiny outcrop. The single red dot just inside the post-perihelion scarp corresponds to two dots in the pre photo. It probably is two in the post photo for two reasons: 

a) if you zoom right in you can see the blob has a sharp hourglass waist betraying the two boulders but they seem to have got closer together perhaps by rolling down the scarp.

b) they must be the same pair because they align perfectly between the fiduciary boulders in both photos. 

Photo 27- the more accurate version of the post-perihelion scarp which is the yellow bull-nose line. 

Photo 28- as photo 27 but with the blue bull-nose line added. These two lines sandwich the width of the ice signature. 

Original for 27/28 but zoomed out. 


Just one paradox remains. We’ve been discussing the erosion across the ice signature (beyond the polygon pattern perimeter) as being from the blue bull nose to the yellow bull nose. This is a westward direction but we know for sure that the overwhelming direction of the erosion front in basin B in 2015 was eastward. The photos at the top of this post prove this. However, in the first available photo, the erosion was well underway. The ice signature is just a narrow strip running north-south. Putting all the evidence together, it would be a reasonable working hypothesis that the erosion started at the blue bull nose. The erosion would then have progressed eastwards across the polygon pattern and also that small distance westward across the ice signature. 

This is in fact entirely in keeping with the findings in Groussin et al. 2015. The authors state that the typical behaviour of the five areas, A to E, is that the erosion started from ice signatures that were located at scarps or rocky cliffs. Area A started from a scarp and went predominantly east before breaking the opposite way through the scarp and going a small way to the west. B started from an east-facing cliff and so it had to go east. C launched east from a shallow scarp but then also went west. D formed its own scarp by burying down, its progress unknown due to having only one photo in the paper. E went a small distance north from a shallow scarp and merged with A. So the erosion at accumulation basin B could have started at the blue bull nose, worked its way east and at the same time eaten through the ice signature in a westward direction for just that short width of the signature. The predominance of east-facing scarps being major players in instigating erosion is related to the sun rising in the east and irradiating these scarps for most of the morning. 

The blue bull nose, defining the back (eastern) edge of the ice signature, may have been a very shallow slope (not a scarp). It would therefore have had to self-excavate up to a point in order to expose the ice signature. Self-excavating would simply mean collapsing due to ongoing sub-surface sublimation and thus sliding down the side of the ice signature. This would automatically reveal the formerly buried side of the ice signature as an icy scarp. This would represent the blue bull nose line extended down in 3D to form an east-facing scarp. This is similar to what happened at area A. Although that area started eroding at a scarp, there was no discernible ice signature on the surface before the erosion started. Area A had to self-excavate to reveal the buried ice signature below it. It started as a tiny pit. Only then could it go east and eventually, west. 


The idea that sublimating gases were coming through the polygon pattern area in previous perihelions would also perhaps explain the polygons themselves. They would be self organising dips in the dust brought about the passage of the gases through the surface dust. This would be rather like so many such patterns on the Earth such as the dimples in melting snow surfaces, themselves self-organising polygons. The long curves formed by multiple polygons would therefore probably betray the shape of the strata bed below and be a result of the slightly differential sublimation vapour pressures coming through the dust directly above those curved fracture planes. 

If you recall, the fuchsia rock curve is actually visible pre-perihelion. So if we had conceptually linked the fuchsia rock curve to the polygon pattern, we might just about have made the leap of intuition that the pattern would prefigure future erosion. We could have made a tentative prediction that there would be erosion all the way to the back of the pattern (or just beyond). This is indeed what happened and it’s much easier in retrospect to see the relationship between the pattern and the erosion. The main smoking gun is the bull-nose curve matching the post-perihelion scarp just beyond it so well. This is then reinforced by the concentric curve pattern behind it. Other patterns elsewhere on 67P might be rectangular or an irregular shape so we wouldn’t be looking for curved or bull-nosed perimeters, like here in basin B, but straight line edges or long wavy lines and these would perhaps betray the extent of future erosion. 

However, what would be common to all these areas is the individual polygons making up the pattern. The polygons seem to betray a weakness of the the whole area in the sense that it’s already weakened by previous outgassing coming through the surface. So if accumulation basin B can be relied upon as a precedent, the polygons making up any such pattern show that these areas are weakened enough to be susceptible to future collapse. 

The relationship between the lines of polygons and the future sublimation was difficult to predict in advance for this primary instance at accumulation basin B. However, it might now be possible to use it in future analysis of other bed layers to predict the occurrence of erosion. This would be the prediction not only of the area that would be susceptible to erosion (i.e. the pattern area itself) but also the line along which the erosion might come to a halt. That line might be the perimeter of the pattern or perhaps one of the curves (or serried straight lines) within that pattern. It might even be just beyond the pattern whilst maintaining the same perimeter shape, as for accumulation basin B in this post. 


1) Temporal morphological changes in the Imhotep region of comet 67P/ Churyumov-Gerasimenko. Groussin et al. 2015. This is free to download:

2) Geomorphology of the Imhotep region on comet 67P/ Churyumov-Gerasimenko from OSIRIS observations. Auger et al. 2015. This is also free:



Copyright ESA/Rosetta/NAVCAM – CC BY-SA IGO 3.0

To view a copy of this licence please visit:



All dotted annotations by A.Cooper