Philae spotted after first landing on 67P

On Friday, we published a series of remarkable NAVCAM images acquired by Rosetta as the orbiter monitored the intended landing point of Philae on 12 November from its orbit above Comet 67P/C-G. The images show what appears to be the shadow of a dust cloud kicked up when Philae made its first touch down on the surface of the comet at 15:35 UTC.

The images were provided by ESA’s Flight Dynamics team on Friday, who sent them to the web team to be published as soon as possible: we all wanted to make sure you saw them right away. Once we had convinced ourselves that the dark ‘splotch’ seen in the second of the images was almost certainly the tell-tale signature of Philae’s precise first touch-down and bounce, we put them out.

However, some careful work by a number of people in ESA’s Flight Dynamics team and by followers of our Rosetta blog has shown that these NAVCAM images show more, namely Philae itself, just after the bounce!


Philae touchdown site seen by Rosetta’s navigation camera. The first image in this sequence was taken on 12 November at 15:30 UTC, just before the lander’s first touchdown; the second image was taken at 15:35 UTC, right after touchdown. The large red circle indicates the position of the shadow of the dust cloud caused by the landing. The third image in the sequence is the same as the second, with the likely position of Philae and its shadow highlighted. Credits: ESA/Rosetta/NAVCAM – CC BY-SA IGO 3.0; pre-processed by Mikel Catania

It appears as a couple of brighter pixels closely accompanied by its shadow in the form of a couple of darker ones just below, both to the right of the diffuse dust cloud shadow.

Credit for the first discovery goes to Gabriele Bellei, from the interplanetary division of Flight Dynamics, who spent hours searching the NAVCAM images for evidence of the landing.

Once the images were published, blog reader John Broughton posted a comment to report that he had spotted the lander in them (thank you, John). There was also quite some speculation by Rosetta blog readers in the comments section, wondering which features might be attributable to the lander. Martin Esser, Henning, Kasuha and Haring in particular were among the first to make insightful observations on the topic, although many others have since joined in.

Last but not least, a careful independent review of the images was made by Mikel Catania from the earth observation division of Flight Dynamics, with the same conclusion. He also made the annotated animation shown here.

So very well done, Gabriele and Mikel from Flight Dynamics, as well as John, Martin, Henning, Kasuha and everyone else from the blog: thank you from all of us for helping spot Philae!

Philae lander operations are done by the Lander Control Centre of DLR in Cologne.

P.S. Note that in our rush to share these exciting images with you on Friday, we made a couple of small errors in the labels and text; those have been updated now. In particular, the image that was indicated to have been taken at 15:30:32 UTC, just before touchdown, was actually an image taken at 16:30:32 UTC, about an hour after touchdown. This should indeed have been clear from the obviously changed illumination conditions, but the web team were very tired at the time.

Also, Flight Dynamics have clarified that the green square on Friday’s images is actually the computed touchdown point based on data taken during the deployment and descent to the surface, not the originally intended landing point. Nevertheless, it’s still very close to the latter, again a testament to the excellent work done by Flight Dynamics.



  • lallo says:

    The dust movement is not too big for a landing in that conditions (with very Low gravity)? The big spot was caused by the rotation of the lander? The harpoon wasn’t surely activated?

    • oneaty says:

      Good points, lallo, In fact, should the lander rotate during the landing? How could a rotating lander succeed in harpooning? Didn’t the rotation contribute to its bouncing (at least, giving it an extra impulse)?

    • oneaty says:

      Also, it’s not completely fool to suppose that the solar panels may have been filled with dust…

  • Any deployable mirros aboard Rosetta?
    perhaps could be used to focus sunlight onto Philae location?

    • Anderson J. says:

      Very clever comment. My first name is Anderson. I love that you thought well outside the box/light source.

    • Oskar S. says:

      The solar panels. Might be hard work to make them continuously reflect sun light directly to Philae while Rosetta itself is in orbit motion around the comet. Would this shed enough light on Philae to provide energy?

    • Chetan says:

      Good thinking

  • John Broughton says:

    I urge the Rossetta team to consider making the closest possible passes over the site to take high-resolution images of the lander and the bounce locations, before comet activity rules that out. Images taken from pairs of points along the path would give 3D information on the lander’s orientation and its surroundings. Likewise, 3D images of the depressions made by the footpads would give clues to the soil density there.

    By the way, the previous blog isn’t strictly in chronological order – I pointed out the exact location of the lander and its shadow at 00:56, earlier than the three credited with spotting it first!

    • Daniel says:

      Hi John: Thanks for pointing out your earlier blog comment, at 00:56 on Saturday. Apologies that we didn’t have a chance to read it in detail in light of what later came to be realised. We’ll add your name to the second post together with the other “discoverers.” Thanks for keeping a keen eye out on our blog!

      • Mesut says:

        Actually its Haring I suppose. At 00:29 he located Philea relative to the boulder,
        “There’s a white smooth spot somewhat 10 pixels east-southeast of the boulder close to the circle (and a dark smooth spot few pixels below it – a shadow?),

      • Haring says:

        You missed my comment on Saturday at 00:23 here:

        I really was the fastest trying to inform you about the white & shadow spot!

        “Haring says: 15/11/2014 at 00:23
        There’s a white smooth spot somewhat 10 pixels east-southeast of the boulder close to the circle (and a dark smooth spot few pixels below it – a shadow?), that could be something like 25m away from the contact point, a plausible distance after a 1’26” rebound. Maybe just a random texture variation… or maybe…”

        No problem, now what’s important is finding our friend Philae directly on the comet!!

        • Daniel says:

          Hi Haring: Nicely spotted! Same as with John’s catch; will add your name to the second post! Thanks for watching — so closely — progress of the mission!

          • Haring says:

            Thank you for the quick reply, and also to Mesut.
            All the compliments to you that made available such a high quality level of information to the public, while keeping alive the discussion. Also in this sense Rosetta mission is *unique*.

  • lesto says:

    the shadow is so away from phylae because it is in its “long jump”? so that mean this is not where phylae is now, right?

    • Steve Jones says:

      That’s my understanding too. We still don’t know exactly where it is, as far as I know.

      • Alberto says:

        You know the exact time Philae was released, the exact time the pictures were taken, Flight Dynamics CAN CALCULATE where the SECOND touchdown may have been !!!!!

  • Robin Sherman says:

    Great news Claudia. The direction Philae bounced would seem to tie up with the CONSERT data placing her on the far side of the landing site B crater. I know some tried to estimate the trajectory from these NAVCAM images, and I guess the flight dynamics team with better access as well as other data are trying to do the same. As I said in a previous post there are plenty of willing helpers here who are happy to scan more images to help the hard pressed teams.

  • Om says:

    Black and white Shining spots appear to be illuminated boulders getting illuminated by the sun light reflected from the Philae .

  • Why so many changes in crater positions between images

    • Darian Rachal says:

      Eduardo, I believe what you are referring to as craters may just be ‘noise’ in the image.

    • Nick in Wales says:

      Eduardo, if you are referring to the white and black dots that are leaping about I suspect they may be noise pixels. Well done to everyone, that is very well spotted!

    • Kappa says:

      These pairs of pixels, which can be seen in the picture, are not actually craters. They might be just data, that were not interpreted as good as the rest and because of the resolution, they are quite visible.

    • Mark Seibold says:

      @Eduardo Pulver – I thought the same, then realized the pixel spots we see change from frame to frame are probably due to ambient visual noise in the process. I am sure we’ll hear an official explanation for it eventually.

    • Ron says:

      Eduardo. You are right. The two proceeding imgages show indeed changing crater position. I guess a crater cannot be erased in such a short period. Perhaps Claudia can explain this in a more scientific way.

    • cheapopete says:

      Probably due to a combination the comet’s rotation and Rosetts’s orbit.

    • aegir says:

      Because pictures are taken from Rosetta wich is orbiting around the comet…

    • Peter says:

      Looks like the angle to the sun changes between shots, changing the peaks that are illuminated. A bunch of “tips” go out and some craters get lit up, disappearing. I’d really want to see a “before” at the same angle to be sure what’s actually different here.

    • Paul A. says:

      The craters are all there. The angle of the sun has changed to make the shadows deeper.

    • David McD says:

      Thanks for asking that question. I had to go back to the images to see what you are talking about. Looking forward to the answer

    • Kasuha says:

      Differences between the two images include:

      1/ different imaging artifacts. When a particle of cosmic radiation passes through the camera’s CCD (or whatever electronic device is used to get the image), it disturbs electric charge in it – apparently in case of OSIRIS cameras, removing some charge from one pixel and adding some charge to another pixel above or below it. These domino pairs are noise and even their average brightness may not correspond to undisturbed image.

      2/ different Rosetta’s position in its travel around the comet. Rosetta is in “orbit” around the comet, so it is moving relative to both its center, and to its surface. Therefore the two images were taken from different angles and the surface is distorted differently due to perspective.

      3/ different comet’s “daytime”. The comet is rotating relative to Sun. The time passed between the two shots means the illumination of the place changed and shadows are differently long and aiming in different directions.

      4/ effects of Philae’s landing, of course.

      Using images posted in this blog I created the following match:

      I edited out most of the “domino” noise, changed overall brightness and used perspective distortion to match surface features (spent WAY too much time on it) but some effects of both Rosetta movement and comet’s rotation are still there. For instance while the surface appears to move fraction of pixel down, top of the boulder on the right appears to be moving up. There’s also difference in shadow lenghts but it is not that apparent.

      • logan says:

        Great work on so little, Kasuha 🙂

      • Robin Sherman says:

        Nicely done Kasuha. This makes it very clear that that bright smudge and shadow are more than likely Philae. Initially I was not as confident as you and others to post that this was possibly Philae, but I am convinced now.

        • Peter says:

          Yes, thanks, that really does make things clearer. And encourages speculation about a “final” resting place – perhaps in the nearest trough to the right.along the trajectory from the first to second impacts.

  • Nick in Wales says:

    What sort of scale is that image? If you have the exact time of the first bounce and Philae’s probable acceleration can you estimate Philae’s altitude in that second image?

  • Bill says:

    Wonderful! I saw several “encouraging spots” on the touchdown Navcams, but given the high emotions of the moment, I feared that pareidolia might have set in.

    Kudos to Gabriele and the Flight Dynamics team for their brilliant performance.

    Do we have an idea of the height of Philae in the second image? Or what the angular elevation of the Sun was at that moment? This will give us a good vector (speed and direction) on Philae’s motion and a better idea of where the poor baby ended up. Of course, FD already has that figured out, but this’ll give us amateur boffins something to twitter over. 🙂


    • Bill says:

      “Black and white Shining spots appear to be illuminated boulders getting illuminated by the sun light reflected from the Philae ”

      “Why so many changes in crater positions between images”


      Those “black-and’white 2-pixel” arrays are simply image noise from the Navcam. They fooled me a couple of months ago into thinking there were several new caters or depressions.


  • Rob says:

    The many groups of two pixels, one dark and one bright, in these images are not craters or moving objects. They are either noise from the imaging sensors on board Rosetta, or artifacts of image processing. So they can be ignored.

    All “real” objects on. or near to, the surface of the comet in this and similar images are composed of at least a few pixels in a “smeared-out” group. Note, for instance, the light and dark “blobs” of pixels in this image identified as Philae and it’s shadow.

  • Robin Sherman says:

    I have been looking at the ROLIS image taken from 40m above the touchdown and comparing that to the NAVCAM image from 15:32 just after touchdown. In the ROLIS image there is a clear fuzzy patch on the image that I interpret as being an active vent. This “active” plume of gas and dust is about 15m from the large boulder at about 8 O’Clock. The centre of the suspected landing point for Philae is also about 14 to 15m at 8 O’Clock from the rock. One would think that the impact of each foot of the lander might create small individual dust clouds in neat triangle, but the cloud of dust deposited has a bulge on one side of the triangle as if gas carried dust in the plume has been deflected by hitting Philae.

    So did Philae in fact land on top of this admittedly small active vent and get propelled up into the vacuum by this gas plume before the ice drills had time to anchor Philae’s feet? Did the dust hitting the bottom of Philae somehow jam the harpoons so the explosive gases from the firing charges actually helped to fire her back away from the surface? Just speculation on my part of course and I don’t know if the team would be able to determine if this was the case, but is this a possible scenario being considered?

    • Tom Tessier says:

      Have we heard details about the harpoon? For example, what signals are needed to trigger it, what sort of mechanism it is (any diagrams?), whether telemetry shows the conditions for the “fire” signal were met, and if the “fire” signal was sent? If there is telemetry giving a continuity check to firing circuits either continuously or a quick test before release from baby’s Mother Ship?

      I am very curious about how it works, not at all critical. Mechanisms are very tough to get exactly right in space under even ideal conditions, and these mechanisms have been in a hard vacuum and DEEP freeze for 10 years!! The fact our baby even separated, landed and worked so well is a grand triumph for the engineers and indeed, a milestone for humanity!

  • Nick Fielibert says:

    The black dots are not craters, but artifacts in the picture

  • Nick Fielibert says:

    The black spots in the image are not craters but pixel artefacts..

  • blake says:

    Surely that is not were philae is now???…

  • FM says:

    @OM, Eduardo Pulver : The white/black pixels that change position from one frame to the other are probably cosmic rays impacts on the detector. Usually those defects are fully corrected before publication, but in this case I suspect they did not have enough time to process the images completely before releasing them.

  • Christian R says:

    Great news, Claudia. Now that the lander’s movement after rebouncing could be described much better I am very curious what that means for the search for the second and final touchdown locations. Please keep us updated!

  • Vladimir says:

    The lander could move approximately to this area:
    It is about 1 km away from first touchdown.
    The following image shows bright spot in the rock shadow.
    The images is a part of panorama on

  • Baznr says:

    Direction and distance, quide to this position:

  • Nano05 says:

    Hey folks first wanted to congratulate the team on such an amazing achievement.

    I also just wanted to reiterate comments from Eduardo above on the vast chages in bolder/crater positions between photos in this sequence. I don’t buy the explanation related to reflections from Philae. I believe the craft too small and too far away from these objects to reflect enough light to make a difference. Also the changes appear in all directions from the crafts relative position. Weird. If anyone else has an explanation it would be interesting to hear it.

    Congrats again to all involved in this historic landing.

  • Steven says:

    Don’t tell me I am the only one curious to see the picture taken at 15:40 UTC :-).

  • Marcelo Malheiros says:

    Can’t ESA use this image combined with the knowledge that Philae bounced another time and the comet dynamics to narrow down the possible area where Philae is?

    I am no expert but it seems more than enough data to run a simulation and find Philae… although I don’t know how that would be relevant at this point…

  • Steve Brightman says:

    I noticed that too in Friday’s image. It also seems that in each case we have a vertically stacked “pixel pair” with a light pixel and a dark pixel beneath it. So I’m wondering if perhaps these are just some sort of imagining artifacts and not actual features?
    Perhaps someone on the team can enlighten us…

  • ysysys says:

    The black and white pixel combinations are artefacts not craters.

  • Janne says:

    From an amaturs point of view 1m/s in denending speed seems very much and it’s pretty clear it will bouns around if harpoon fails. The gravity is very weak but still it would have accelerated toward the sufrace hitting the “ground” pretty fast probably up to 3-4m/s. I just would have made passive bounce free landing gears like volvo crash-colapsing seats.

  • Tom Bechtold says:

    Is that Philae that appears top-left immediately outside the red circle in the 3-frame sequence posted 16/11/2014?

  • Vladimir says:

    I’ve tried to make more accurate image combining. I’d searched the lander approximately in this direction.

    • Gabriele says:

      Very good job! I think that if Philae is landed in the crater we can see in your image Rosetta won’t be able to see and locate it. Too much rocks and shadows.

    • galahad says:

      Nice work, Vladimir, but I think the final landing site of Philae might be far off that straight line if you take into account the rotation of the comet (the fist bounce took quite a long time, over an hour if I’m not mistaken, and I bet the plane of the bounce trajectory was not exactly parallel with the equator of the comet).

    • David G. says:

      Vladimir, since the comet was rotating beneath Philae during the bounce, I wonder if it is valid to assume that the trajectory is along the line that you have drawn, which is defined by the first landing location and the location of Philae 86 seconds into the bounce. Another way of stating this is that it may not be valid to assume that Philae was moving in the same frame of reference as the surface of the comet.

      If my thinking is correct, then the “ground track” of the bounce would have to take into account the motion of the comet during during the bounce. One visual aid to this problem is this animation:
      54 seconds into the video, Philae lands on 67P, then one can imagine the bounce happening while the comet continues to rotate in space.

      Returning to the image of the landing site, It is hard to visualize, but I think that all of the surface features in the image would move parallel to the comet’s equator. If you made this transformation and then added it to your trajectory line, you would get the final resting place of Philae on the surface.

      Does any of this make sense, or are my assumptions incorrect?

    • Cometstalker says:

      Hi Vladimir, in general good start but try to add the Coriolis effect as the comet spins fast and the jump is a long time. Your straight line will then be curved in the horizontal projection. The trajectory is very flat making the distance quit long. The good news is that any air resistance is not present if you calculate the ballistics. The bad news are that the gravity vector is changing a lot during this flight. All in all a tuff calculation and an easy estimate is that the new location is not “J” and definitely needs a new name, find it and call it as you please as this is your right.

  • I. Newton says:

    Now the time of flight of the first rebounce of the lander is known, it is possible to estimate the height of the rebounce and the vertical velocity with which the lander was rebounced, using some simple Newtonian math.
    It can be shown that the height, in terms of the comets gravity constant gc and time of flight t, can be expressed as:
    height =1/8*gc*t*t
    t= 1 hour, 51 minutes = 6660 seconds.
    According to the Rosetta lead scientist the comets gravity constant is approximately:
    gc= 0.00001*g_earth = 0.0001 (approx.)
    This yields:
    height= 554 meters
    For these calculations it has been assumed that the gravity constant is approximately independent of height, which is valid only for small height, but probably not far off.
    The vertical rebounce velocity can be expressed as:
    V= 0.5*gc*t
    which yields
    V=0.333 meters/sec
    This is surprisingly high, given that the landing gear was designed to absorb most of the kinetic energy of the lander at touch down

    • Fig Newton says:

      You’re assumptions are based upon a spherical body of uniform mass. This comet is neither spherical nor uniform. There are many other factors involved as well, such as mass of the two bodies, angular velocity on approach, touchdown contact orientation, localized gas venting from the comet’s surface and more! However, thanks for stopping by and have a great day!

    • Steven says:

      I agree that reality is more complex. However I found a hight twice yours. If I remember correctly that is what esa communicated. 1 km high and rebound speed of 0.33 m/s.

      • amieres says:

        If I remember the announcement correctly, from the landing gear, they measured a rebound speed of 38 cm/s. That must have been the whole vector including horizontal speed.

        I think the formula I. Newton used for the height is incorrect. It should be 1/4 and not 1/8, that would explain why you got a different result.

        Also for such a small object gravity changes a lot depending on distance. If the first touchdown place was, lets say, 2.5km from the comet’s center of mass and the lander rebounded 1km, at that point (3.5km) gravity would be about half of what it was at the surface. That shows that gravity cannot be assumed to be constant, instead it reduces by the square of the distance from center of mass. Which means elevation would be even higher.

        As interesting as calculating the bounce height is, I don’t think is relevant for calculating the second landing. With the picture of the lander we have an estimated horizontal velocity and the time of second touchdown, that should be enough to calculate the horizontal travel distance:

        The lander travelled 22 pixels in 86 seconds, then in 6660 seconds it would have travelled 1703 pixels, at 1.3m per pixel that is about 2200 meters.

        (The term ‘horizontal’, is just a matter of speaking, the circumferential distance would be a more appropriate term)

        Since the comet is not spherical but very irregular, what would be needed is a section of the comet showing the contour of the surface in the direction the lander was travelling to estimate its second touchdown place.

        • Max Beran says:

          It’s worse than that arnieres. The net force on an object due to gravity is the superposition of the gravitational attraction of every mass element of the attracting object. This is well approximated by using the centre of mass of a regular distant object but not in the case of a close by irregular object like 67p. I think in this case the effect of “smearing” would be to reduce somewhat the vertical change in g because the reduction is more marked at lower than at higher elevation.

          I don’t know how the team solve this – perhaps armed with a supercomputer they can actually do the superposition. Maybe they have some semi-analytical approach approximating the comet’s shape. Oh, and then there’s the effect of density – that also enters into the gravity formula for each volumetric “pixel”. I think they have a figure for the total mass but doubt we know much about how density varies within the mass.

  • Nicolas Papageorgiou says:

    How can we explain that we can observe a very distinct shadow of Philae while at the same time it’s said that her solar panels are barely exposed to the sun? Is it due to the orientation of Pfilae herself, such that the panels are not facing the sun?

    • Ramashalanka says:

      The 2nd (and 3rd) images of the gif are just after the first touchdown. Philae has bounced and is (just) in the air. The dust cloud is (near) where it bounced. The final resting place is some out of the frame.

  • Vladimi says:

    I. Newton, your touchdown velocity is close to real 0.38 m/s. Given g = 0.0001 m2/s and time of first leap 1:50 we have approximately 1.2 km. With resolution of 1.3km/px we have the following area of possible second touchdown.

    • Vladimi says:

      Sorry, 1.3 m/px.

      • Haring says:

        Vladimi, you did a very good job but the distance in your graphical photo is much less than 1.2 km, is rather something in the order of 200 m.
        Just consider that Philae traveled something like 25m in just 1’26” (the distance between the small cloud and the small spot).
        Philae should be much more distant.

        • Haring says:

          …and the boulder next to the Ist touchdown is around 5m diameter (t’s written somewhere in the Esa Rosetta website, some days ago)

    • Christian R says:

      What you show is less than 250m away from first touchdown. Look at this NAVCAM image which covers 2.3 x 2.3 km (1024px with 2.27m/px):

    • Robin Sherman says:

      Great work Vladimi. One issue you have not considered is the rotation of the comet while Philae was above the surface and whether she was still gravitationally bound to the same extent to enable the straight line you have drawn. Unfortunately I don’t know enough about flight dynamics and the direction of rotation with respect to these images to help. It does illustrate to us amateurs the sort of process that the flight dynamics team can use to find Philae, so thanks for your efforts. I am sure that your estimate is not too far out and like you say this is at least a region to start a concentrated search.

      My intuition says Philae’s second touch down may have been at the top of that crater wall, very near to it’s edge and she bounced or toppled over the edge. Such is the low gravity it would take several minutes to reach the bottom, during which time the weight of the landing gear and flywheel would have been enough to get her upright during the fall.

      The quality of the telemetry link suggests to the lander team Philae was pretty much upright. From the position you suggest, one view would have been across the crater to the side where the rim has been eroded and since the crater floor also slopes downwards away from the base of this crater wall, the view would have been of space, as one of the CIVAS images showed. The brilliant mosaic of the two CIVAS images we have seen show cometary material at the base of the “cliff”, but which has fallen a short distance away from the “cliff”. Many of the “Cliffs” we have seen from the NAVCAM images have quite large overhangs so Philae may not have hit the crater wall on the way down at all. Hopefully the CIVAS images taken later on with a better exposure time will enable some distances to the wall behind her to be calculated.

      The MUPUS instrument was deployed but the surface was so hard that the shaft of its hammer broke as reported in the BBC’s “Sky AT Night” special tonight.. The question of, is the surface rock or ice remains an open one, for us at least. The harpoons may have fired as planned, but bounced off the very hard subsurface, the ice screws were deployed but could not dig into the subsurface for the same reason. The plot thickens as they say.

    • doekia says:

      I think you forgot that during 1 hour 50min comet is revolving, hence you cannot draw a straight line to estimate 2nd touch down location

    • Jeremy Robinson says:

      Thanks Vladimir! Greatly appreciate your posts and pics.

      To calculate the distance of the first bounce, we could approach it another way, too:
      If we know the time of first TD, and time of Rosetta photo capturing Philae in first bounce… along with approximate lateral distance traveled between those two times, we could calculate lateral velocity of first bounce. Then multiply by the total bounce time (1:50)….

    • Marcelo Malheiros says:

      Are you considering the comet rotation? With Low gravity… Philae should be doing a curve and not a straight line…

  • Henry says:

    Well done to those observant writers and ESA scientists!

    Vladimir, your bright spot looks very interesting…

  • Nicholas says:

    Your projection lines up with the direction of the supect dust plume. So it could be a possible candidate.

  • Jörg Wagner says:

    I am quite sure that the distance that Philae has covered with its first hop is much bigger than calculated by some commenters here. The math needed is actually quite simple:
    Philae made 22 px in image x direction and 3 in image y direction. With an image resolution of 1.3 m/px this gives us a speed along the surface of ~29 m in 86 seconds (time between touch-down and image taken). As the duration of the first hop is known, we only need this horizontal component of the velocity for a first approximation. In 1:50 hours it results in a flight distance of over 2.2 km.
    I am very aware of the fact that due to the irregular shape of the comet this is by no means a standard ballistic flight path. But as a first approximation it is valid.
    2.2 km would mean 1/3 of a full rotation around the smaller lobe. It would mean that Philae landed somewhere in the neck region or even beyond that on the inner side of the larger lobe.

    • Cometstalker says:

      You are right and to fuzz up the orientation a bit more the trajectory is curved in two planes, one formed by gravity however this is shaped and the other one formed by the Coriolis force that is also not easy to calculate.
      I bet the flight dynamics team has the tools and data needed to estimate a landing spot #2 that is within two acres of accuracy and a landing spot #3 should be close enough.

  • Daniele says:

    This mission is so incredible, i’m always tuned for news and updates! Compliments, you made a great job!!

    Watching at the pictures where Philae bounced you can see a rock very close, it was a big lucky that he didn’t hit the stone!

  • Pedro Garcia Caballero says:

    Hi! I’m search philae too! I think like Vladimi, and if you search in line direction…

    I see this in new picture from osiris:

    Philae are you there??

  • ian mcdonald says:

    Is there anything on rosetta that could be used to reflect sunlight on the landing craft?

  • logan says:

    Great collective work. Please don’t discard another scenarios. As example, the object and shadow being a jet ‘quaked’ by the impact.

    • logan says:

      Weird as it sounds: Are the neck surroundings too far [in time] as to be considered a search area?

    • logan says:

      The second ‘jump’ is not a reboot like the first, too small relative to first, looks more like if Philae where ‘cached’ by the soft material, in a ‘perpendicular’ way, given the window insolation time.

      • Marco says:

        I was thinking bounced off a cliff into a crevice. Not too much soft stuff on the surface. I think the legs absorbed the last of the momentum.

    • logan says:

      The impact line of debris is pointing differently to the object.

  • Daniel says:

    Here is a combination of an OSIRIS image that was posted earlier (In the Rosetta blog post: Three touchdowns for Rosetta’s lander) along with the NAVCAM touchdown image Also a version with the touchdown dust cloud (I guess?) and Philae along with its shadow roughly highlighted at It isn’t a perfect match but I didn’t consider it worth putting any more work on it.
    Looking at these pictures it does seem like it was moving quite quickly. That’s the distance covered in 1 minute and 26 seconds of a trip that took almost two hours after all.
    When speculating on where it ended up, do keep in mind that we aren’t looking at a flat surface, you can’t draw a straight line on the image and expect to find the final landing site.

    • Daniel says:

      Should probably add this:
      Images are a combination of CAMA20141112153532snip_paint.png (Credit: ESA/Rosetta/NAVCAM – CC BY-SA IGO 3.0) and ESA_Rosetta_OSIRIS-NAC_Landing_site_50km.png (Credits: ESA/Rosetta/MPS, for OSIRIS Team MPS/UPD/LAM/IAA/SSO/INTA/UPM/DASP/IDA) that were posted on the ESA Rosetta blog.

  • Gordon Foat says:

    A UV strobe light led would have been worth mounting to each side of the lander. Perhaps you could spot a reflection off the PV cells? Just a thought,…

  • KIRKOUT says:

    Since you have the information regarding the impactor mass and speed, and the dust cloud what can the ejecta tell you about the physical properties of the comet? There must be some sort of a relationship there. The impact cloud seems to have SE component to it spreading many metres even if shadow has a part in this and the fan like shape that you would expect from an oblique strike looks to be present. The linear mound obviously shows. I refer back to earlier picture that gives a good view.

  • wasserball says:

    If you look at the large white object,(near the center of the photo) and right below it is a dark spot , which I assume is its shadow? If that is so, the shadow for second landing spot of Philae looks too far away.

  • Juan Simón says:

    Vladimi, shouldn’t we calculate the comet’s rotation below Philae after rebound to estimate their final landing? There’s no atmosphere to drag lander with rotation…

  • Sage Wiseman says:

    Look at the two pictures. You can clearly see orbiting boulders. As interesting as the lander

  • Kevin Stewart says:

    I admire the work and the achievements of the entire Rosetta/Philae team. I was very happy when I learned that Philae was strong enough to be able to complete its last science block, and was even able to make a change in position with its last bit of power. Now that you have more information about the bounce trajectory, I hope Rosetta can locate Philae in its niche, and that in time Philae may even reawaken and speak to you once more. Bravos all around.

  • Jon says:

    One of these comments reminded me that there is still some hope for Philae. She is untethered, since the harpoons failed. A tiny weight on a tiny rock of dirty, dusty ice, as it approaches the sun there is a reasonable chance that the obstructing cliff, being higher, might go first or that the out-gassing might shift her position giving her sunlight. If so we could be in for the ride of a lifetime! It was never planned for her to record the actual passage of the sun. I do hope the team do not simply give up on her since she is taking a nap! Of course Rosetta will be followed

  • David G. says:

    Others have pointed out that a straigh line trajectory drawn on the image may not be correct. Here are my thoughts on the subject:

    Since the comet was rotating beneath Philae during the bounce, I wonder if it is valid to assume that the trajectory is along the line that you have drawn, which is defined by the first landing location and the location of Philae 86 seconds into the bounce. Another way of stating this is that it may not be valid to assume that Philae was moving in the same frame of reference as the surface of the comet.

    If my thinking is correct, then the “ground track” of the bounce would have to take into account the motion of the comet during during the bounce. One visual aid to this problem is this animation:
    54 seconds into the video, Philae lands on 67P, then one can imagine the bounce happening while the comet continues to rotate in space.

    Returning to the image of the landing site, It is hard to visualize, but I think that all of the surface features in the image would move parallel to the comet’s equator. If you made this transformation and then added it to your trajectory line, you would get the final resting place of Philae on the surface.

    Does any of this make sense, or are my assumptions incorrect?

  • Joachim Jastrow says:

    Hello Vladimir, your line starts in the center of the dust-cloud’s shadow. I suppose the (invisible) point of first touch-down has to be somewhere above the shadow of the raised particles.

  • l.c.b says:

    Je suis très impressionner par cette mission bravo a toutes les personnes qui on participer à cette aventure génial

    Christophe Lefebure.

  • Paul says:

    Interesting trajectory, Vladimi. Please consider the calculated initial touchdown point is a few pixels higher. See green dot in another picture.

  • wg says:

    Two random impressions:

    – while agreeing that the evidence for the postulated original landing site seems reasonable the two pixels speculated to be the final resting site are to the naked eye statistically totally meaningless given the noise in these images on this spacial scale.

    – huge congratulations to the flight dynamics team, exceptional performance. Major boo to the team that designed and programmed the landing sequence. ESA should have enough data on their hand now to explain why the landing/anchoring sequence failed so utterly. The lander apparently did not even realize it hit the surface. The cold gas thruster did not activate, ice screws did not work, nor the harpoon system. Why, what failed? Undue reliance of the whole design on some one critical element, sensor? A software bug? What?

  • Johan Prins says:

    Robin Sherman wrote: “..during which time the weight of the landing gear and flywheel would have been enough to get her upright during the fall.”
    Such thing would probably not happen, unless there was a strong enough gravity gradient, I’m afraid, and the lander would just continue rotating around it’s own center of gravity. Still, we are all *very* lucky that the lander did not land upside down, and that it could do almost everything 😀

  • Christian R says:

    So it seems that the horizontal velocity of the lander after rebouncing was rather high, more that 0.30 m/s. A point on the comet’s surface 2km away from the rotation axis moves 0.27 m/s. If the rotation axis is stable (and I think this was said a few weeks ago) the comet rotates nearly “downwards” in the NAVCAM image (towards the 6 o’clock position). This together with the effect of the gravitational pull on Philae’s flight curve during the 111 minute hop (plus potential other forces I am not aware of) is what the flight dynamics experts will try to bring together. I am really curious where we will end up. Great job!

  • Michiel de Jong says:

    You can actually see the location where Philea has hit the ground! Just ‘north’ of the dust cloud! The cloud casts it’s shadow to the south. Based on this position and timing of the photo’s they must be able to determine speed and direction after the bounch!
    And of course indicates wher it is now!

  • Paul says:

    I have been following this incredible mission closely since rosetta was woken from deep sleep. At this time, it seamed unlikely that you would release any science data close to real time. (I read about the 6 months rule…)

    Now you are sharing data even before taking your own conclusions. This makes us feel like a little explorer, actually taking part in the mission. Yay!

    Thanks to the rosetta team for sharing observations so timely and making us all part of this adventure! This really shows that you did listen to our feedback.

    • sepp says:

      +1 to all of this.
      Thank you, everyone at ESA and the connected agencies!

  • Christian R says:

    It is interesting how Philae’s direction of motion changed due to the rebound. The ROLIS picture taken 40m above ground suggests that the lander came down pretty vertical. The NAVCAM image shows that it bounced off “to the right” pretty fast. If the published velocity of 0.38 m/s holds true the angle between flight curve and surface should have been rather flat, between 30 and 40 degrees. Quite a change of the vector.

    • sepp says:

      True. But don’t forget that the comet rotates pretty quickly!
      So even if the rebound were straight up again, after 2 hours of flight, the comet would have rotated quite a bit.

      It’s still probable that the rebound was not exactly vertical. This could depend on the slope of the terrain at the site. Also, maybe some of the legs hit a harder surface than others, meaning they rebounded more?

  • Bernhard says:

    Great work, it is amazing to see Philae bouncing off the surface. I am sue you will find where Philae is hiding as you now will be able to estimate direction, speed, and trajectory….
    Since days I am wondering if there are no Philae images from the first landing site. Shouldn’t CIVA have stated taking pictures at first touch down? And shouldn’t there be also pictures from its 2h bounce as well? If someone could answer me this question? Can’t wait to see those if they were taken….

  • Cliff Gobreski says:

    Sorry if already addressed; Have Philae engineers pinpointed the exact cause for both harpoon and thruster fail’s? i.e. able to monitor thruster fuel levels before/after 1st impact, electrical status of both firing/initializing circuitry, did either piece receive signal to initiate? Were able to monitor equip. status’ for past 10 yrs? p.s. Those are some rugged cameras.

  • Jacob nielsen says:

    It is predicted that Philae will succomb to high temperatures during the passage of the sun. But what temperatures does that mean? I havent found any actual predictions of the temperature profile of the comet during passage. When Rosetta approached 67p in august, the surface temperature was found to be some 20 degrees warmer than predicted, so I assume no real measurements have previously been carried out? I imagine the temperature rise comes from a combination of solar radiation, a greenhouse effect caused by gasses and vapors of H2O, CO2, CH3 mainly, and a pressure buildup. By now a revised prediction of the temperature development should be at hand at ESA. The temperature chart that was presented months ago should exist in more recent editions.
    Question: the housekeeping data from philae includes ambient temperatures , so how does that correspond with the latest chart, when considering both 2. and final impact?
    Question no. 2: what is the predicted ‘expiry date’ for Philea when considering temperature in the current location? (With a number of assumptions)
    I don’t expect ESA staff to respond to this, I assume numbers are being crunched right now.

    • doekia says:


      Temperature rise for Philae is solely by radiation due the nature of weak to no atmosphere. That said you have the face facing the sun rising rapidly an no – or close to none dissipation because again no atmosphere.

      Perihelion is at 1.2UA – almost the distance moon-sun. Temperature on the moon surface exposed to sun is arround 123 Celsius (223F) while dark side is about minus 153 Celcius ( -242F ). That is roughly the number Philea will have to face. Put that in scale of the fact that the portion in the shadow of the clift will be close to the dark side temperature and enlighted portion close to bright side – add thermal dilatation stress, electronic component facing this by thermal conductivity you get the picture and the explaination why it is say she will barely survive perihilion…

    • logan says:

      Hi Jacob 🙂
      “…, a greenhouse effect caused…”

      Cordially remember you that sometime ago was established that ‘atmospheric’ pressure is not going to be above the millibar. So no greenhouse.

    • logan says:

      To my little mind there are 3 scenario changes that bring the slight hope of Philae survival to Perihelion:

      – He is now sort of surrounded by a hipotetized thermally isolating ‘lakritz’.
      – If ‘Elephantine?’ site turn out to be a ‘pit’ then he is probably down under a ‘deposit’ of cold gas.
      -The lesser insolation window.

  • Stefan says:

    I guess it should be possible to calculate the movement of the shadows, while 67/P is approching the sun. It’s beyond my technical capabilities but I guess models of that have already been made or will be made soon. Is there already a simulation showing the development of the shadows for the next month in the region where Philae is expected to be?

  • David says:

    It would be great to have more images at hand… Kind of “crowdsearching” for Philae. I think this would at least help you discard some obvious and other not so obvious candidates, and would also help gathering a set of spots of interest to look closer.

    Many of us spotted the lander in the second image, although I understand that this official statement required a thorough analysis as it surely has been done before publishing this. So, good job on this too, ESA!

    I’m sure the landed Philae will be much harder to find, but again, there is plenty of people willing to help on this. Probably better if having some directions on what to look for.

    Proud of your astonishing good job!

  • Cometstalker says:

    Depending on the surface properties and the performance of the electrodynamic brake system plus the tension in the tree legs, just about any debounce track is possible, never the less as the trajectory is by some extent known by now the search area is reduced quite a lot. Flight dynamics calculations might reduce this area to within an acre or two. The calculation is complex as the gravity uniformity, the Coriolis effect at this two hour jump and the fuzzy input data gives a great variation. It would be nice to get a 3D animated estimation of a trajectory of this first debounce and an estimated search area where Philae might be now. Help from the Bloggers seems to be sufficient and it would be fare to help the Bloggers as well to make them continue their great work.

  • Tim Putnam says:

    How come the search for Philae has descended into a game of Wheres Wally?

    I’d be fascinated to learn more about how you map and know the position of Rosetta when you’re calculating the orbits, and why you can’t know the relative position of Philae. Are there technologies that were left out that might have helped? What measurements do you know – e.g. can you estimate the straight distance from Rosetta?

    How accurately do you know the position of Rosetta and how do you describe it – is it polar coordinates, and if so, around what?

    • Tim Putnam says:

      I’m also curious to know if the most likely direction of the bounce was knowable when choosing the landing site? Was the way it bounced random – was the rebound velocity or the spinning of the comet during the bounce the largest component in changing position on the surface? If the bounce direction was predictable, with hindsight, would it have been better to choose a landing spot with the maximum chance of finding good terrain after a big bounce since the thruster had failed? Just curious. You’re publishing images (some of the best every published in history not complaining!) but a lot of us would be just as interested to see other kinds of data too.

  • Would be interesting to see IR images from OSIRIS on the landing region. Shouldn’t be Philae slightly hotter than its surroundings ? At least during the first hours after the landing.

    • logan says:

      Smart comment, Jose Luis. Hope the teams are using those to reduce search areas. Don’t forget that spaces drones use to be thermally isolated, but not all of their surface. 🙂

  • Vladimir says:

    Thanks all for attention to my amateur assumptions. When I did some calculations are didn’t consider that I was working with resampled image and actually didn’t really realise its scale. Now I understand that straight line is not applicable, and direction is affected by comet rotation. Second touchdown is much farther than I marked.

    I think wide angle image prodived by Daniel gives more information about direction of the second touchdown.

  • Heiner Wolf says:

    So, Philae bounced before coming to a halt. ‘Bounce’ does not quite describe it correctly. Philae did by far the biggest #jump ever. The tallest, longest, slowest JUMP ever done by a human made object (including humans). Without using thrusters, just by pushing off from a rock. That’s something for the Guiness Book of Records.

  • Max Beran says:

    I am intrigued by the absence of bounces beyond the second. What this suggests to me is that Philae’s flight was arrested by collision with a vertical(ish) obstacle like a cliff face and not, as seems to be the general inference of many, that it came to rest naturally at a location that was by bad luck in the shadow of a cliff!.

    Following through on this tack and using logic similar to I Newton’s above, how about the following? Assume the energy lost during the second landing was a scaled down version of that from the first (which was down from 1 m/s incoming to 0.38 m/s outgoing) then one might speculate that the launch velocity into the second bounce is of the order of 0.144 m/s. Assuming again that the second bounce was otherwise elastic it took off at about 69 deg to horizontal (perpendicular to line to comet’s centre of mass). Then in 7 minutes flight time with g=.-0.00017 m/s/s the position of the lander would be c. 20 m from the bounce horizontally and 40-ish m above. Are there cliffs of this height around?

    I should explain that my calculations are based on a dumbbell-shaped comet and some round number guesstimates of the positions of centres of mass and relative masses of the comet’s two lobes. I also assume the action takes place along the line of the dumbbell. The requirement to fix the flight time of the first bounce is quite constraining in this toy version of reality as far as implied g and angle etc are concerned – hence the funny numbers like -.00017 and 69.

    So we need some oblique shots of cliff faces with poor Philae dangling helplessly part way up.

  • Bram says:

    A rough back-of-the-envelope calculation: comet radius ~ 2 km, rotation period ~12h leads to a rotational speed at the “equator” of 0.3 m/s, in the same order of magnitude as the apparent vertical speed of philae after first touch-down. So if the direction of philae is along the rotational direction of the comet, it may be possible the main vertical velocity component of philae is due to comet rotation.

  • Gilbert says:

    It would be interesting to see later images of the first landing area. If the shadow of the dust cloud is still there – that might be OK, but if Philea & its shadow are still there then its back to square one.

  • Darragh Owens says:

    Congratulations to the Rosetta/Philea team for this visionary, inspiring and daring scientific adventure. It is already more than a huge success. Any further results will be a golden bonus.

    Following the story like a fast-paced thriller!


  • amieres says:

    Based on my estimations this would be roughly the spot for the second landing:

    Closer view of estimated 2nd landing.— amieres (@amieres123) November 17, 2014

    • amieres says:

      This is based on an estimated surface trajectory of 2200 meters from the first landing point.

      Calculated like this:

      22 pixels * 6660 sec / 86 sec * 1.3 m/pixel = 2215

  • Oliver says:

    If the sun doesn’t shine directly to Philae, there is perhaps -in principal- another way to load the batteries:
    If Rosetta can be taken nearer to Philae, it can deflect sunlight to the Lander from its solar panel.
    The preconditions are, that the position of Philae is already found and Rosetta can be controlled in such to hold the beam long enough on Philae

  • JOHN says:

    Doesn’t the shadow allow a good estimate of altitude and an even more accurate estimate of the trajectory through space than just horizontal velocity?

    • Max Beran says:

      Fixing the trajectory is fraught with difficulty due to the variable nature of the gravity field changing with height and lateral position. And the second bounce is even harder than the first as we don’t know how much momentum was lost in the bounce (nearly two-thirds was lost at the first).

      I reckon the ground speed quoted by arnieres is much too fast as this cannot exceed 0.38 cos(theta) m/s which was set at the first rebound. (theta is the post first bounce angle)..

      However your expectation is correct – knowing the vertical and horizontal components of the trajectory will help to fix the trajectory better than the horizontal alone and would help to constrain the local value of g.

      Even the notion of “horizontal” is tricky. On Earth it would be the plane perpendicular to the line leading to the centre of the planet. But on a dumbbell shaped object…

      I still find it highly unlikely that the second bounce completed, much more likely is that it was interrupted by Philae being stopped by hitting up against a cliff (see my earlier post for guesstimates of horizontal and vertical displacement). I am surprised not to see this much more probable scenario spelled out by the team.

  • JerryN says:

    As the comet becomes more active and begins releasing large clouds of material, I wonder if that material might reflect sunlight into the shadows and thus help Philae charge its batteries…

  • Martin Schäfer says:

    Did the harpoons definitely not fire or is this statement a conclusion of the effect that philae bounced?

    How does the shape of the harpoons look like?

    Is it possible, that the harpoons were fired, but could not anchor to the ground and rebounced themselves. That could possibly explain the rotation and reinforce the bouncing of Philae.

    • Laurent Chabin says:

      Missions managers said sensors indicated they did not fire. You can see the harpons , search for Documentary HD – Rosetta Mission – Landing On A Comet on youtube. Managers also said that during tests on earth of the harpoons pyrotechnics, they had had failures, after many years in stock.

  • can-rnb says:

    The whole operation is so amazing. Congratulations to the whole Rosetta team for this interesting and visionary scientific journey!

    I enjoyed following the entire story.

  • Guy Gibson says:

    Everyone seems to be looking compleatly in the wrong direction and miscalculating the speed after the first bounce. The area that ESA was at first looking at ,was following “in a straight line” Philaes landing path but as it´s now clear in the photos of the descent that the lander on bouncing, changed direction by 43º approx. The speed, by mesuring the distances against the times of the descent photo sequence, the distance travelled after bouncing (15:34:06) and the subsequent photo, (15:43:?) highlighting it over the crater shadow, 8minutes and 54 seconds later. The distance travelled is only a fraction less than the descent speed, in some of my calculations it even gained speed!!! Maybe the comets rotation landed on Philae rather than Philae landing on the comet.

  • flug says:

    FYI after the discussion here yesterday I got messing around with the Rosetta/Philae/Comet 67P simulation in Orbiter Space Flight Simulator.

    You might be interested in taking a look at the results. It has two interesting results:

    – A pretty good indication of where Philae landed (near the top of the giant cliff/hill on the far side of the neck)

    – Comet 67P 3-D mesh imported into Orbiter, which gives a nice way of visualizing Philae’s movements against a pretty realistic background.

    Take a look here:

    Couple more quick simulation runs here:

  • I watched Neil Armtrong step on the moon on a small tv in West Berlin as a child (dad was stationed in the Britiish army there).

    Now I am watching a robot land on a comet, an equally amazing feat, well done everyone. Philae please phone home…

  • I am waiting expectantly for ESA to narrow down the final resting place for Philae. I am confident they will find her.

  • Luca says:

    Take a look to this amazing video with simulation of the two bounce of philae :

    Hope we will have some news from esa…any kind of news.

  • Luca says:

    also here, some interesting simulation about position of philae

  • Lots of information here! Thanks!

  • Martín Gil says:

    Quisiera saber , al final , que nombre le asignaron al sitio de aterrizaje de Philae, quien gano el concurso. Gracias

  • Image sensor says:

    Great writing it is such a good and nice idea thanks for sharing your article .I like your post.

Comments are closed.