(Will be updated as project proceeds. Content by James Whitby, with contributions from Leandro von Werra, Prof. Nicolas Thomas, Dr Veerle Sterken, Dr Frank Preusker, Dr Frank Scholten, Raphael Marschall)


Science themes addressed by the MiARD project


Relief maps

The LOMID project seeks to combine two different techniques which can be used to generate a digital terrain model or relief map from a set of images. The two techniques are

  1. SPG = stereophotogrammetry. This is the use of stereo pairs of images of the same area taken from two slightly different locations to deduce heights (essentially the same way our brain infers depth from our two eyes), and is the ‘classical’ technique used by cartographers to generate relief maps from aerial or satellite photography.
  2. SPC = stereo photoclinometry. Instead of using images taken from two different locations (angles) this uses images of the same area under different illumination angles. All other things being equal, the brightness is then related to the slope.

SPC is better than SPG for determining relief due to rather subtle slopes, whereas SPG gives better results for rough areas. SPC can give a higher resolution model than SPG because information even for single pixels is obtained, but the SPC approach needs additional constraints which in our approach come from a lower resolution SPG model. The MiARD project will create the most accurate possible three dimensional representation of the comet’s surface by combining results from both the SPC and SPG techniques, using the thousands of images taken of the comet during the ROSETA mission. Accurate maps of the surface, from different locations (times) in the comet’s orbit around the sun will show the changes in the surface due to the loss of material as the comet heats up and evaporates when it approaches the sun. The particular approach to SPC taken by the MiARD project is known as Multi-resolution Stereo-PhotoClinometry by Deformation (MSPCD).

There are several challenges to be overcome by the project members in creating the digital terrain model (sometimes also simply known as a shape model of the comet).

  1. The comets surface is changing all of the time. This makes it difficult to use image pairs consisting of images taken at different times, and, together with a lack of many images of some areas such as the southern hemisphere, makes it difficult to simply define pre- and post-helion shape models.
  2. Because of the rotation of the comet and the Rosetta spacecraft, overlapping images are often taken from different distances, with different orientations and illuminations. This makes the stereo reconstruction process much more difficult than for say an aerial survey on Earth.
  3. The coordinate system of the comet is complicated because it’s surface is not well approximated by a sphere or even an ellipsoide.
  4. There is a huge amount of data to process – there are tens of thousands of pictures of the surface, and the resulting final shape models have fifty million points (intermediate calculations use over a billion points!)
  5. Assessing the quality (accuracy) of the shape model

Geological maps

Using the nature of the surface (smooth or rough, craters or cracks…) and spectroscopic information about the composition, ‘geological’ maps of the comets surface will be made, grouping characteristic surface types together.


The dust particles emitted by comets are moving at high velocities compared to some other solar system objects, and so pose a potential hazard to satellites and spacecraft (e.g. the Perseid meteor shower is thought to be due to debris from the comet Swift-Tuttle which has a period of 135 years). At the moment, the exact nature of the dust and so the forces acting upon dust particles are not known accurately and so trajectories are not certain. Although dust particles such as those responsible for the Perseid meteor storm are usually much less than 1 gram in weight, they are moving at about 58 km/s with respect to the Earth, so they have a lot of energy – comparable to that of a rifle bullet. When Perseids hit the moon, they cause an explosion at the surface that can be seen from Earth.

Although meteorite or dust induced damage to satellites is rare, it is likely that the OLYMPUS communications satellite was badly damaged by a Perseid (due to a temporary electronics problem that used up too much fuel for attitude control) during the unusually intense 1993 pass. It is known from measurements that other satellites (Pegasus satellites equipped to measure such impacts) have been impacted. The risk to satellites from meteor storms is discussed in this peer-reviewed paper from 1997 (no open access), in which it is estimated that for every 50 meters-squared of surface, there is a less than 1% risk of being struck during an average meteor storm.

Cometary tail

When we see the ‘tail’ of a comet, we are mostly seeing sunlight scattered by dust particles. If you look closely at a comet from Earth, you will see that the visible tail is split into two – one of these consists of dust particles (made visible by scattered sunlight) and the other consists of ionised gas molecules (also made visible by sunlight, but by resonant scattering at certain wavelengths only).

Effects on comet’s trajectory

Dust particle trajectories

Cometary outgassing and material loss i.e. ‘activity’

In order to interpret measurements of the cometary surface and tail, we need to understand the processes that occur at the surface that result in the comet losing material each time it approaches the sun and is heated. Even before we had close-up views of comets, we knew that they could not be pure ‘snowballs’ of frozen water and frozen carbon dioxide – a simple but robust model of such a snowball predicts mass loss rates at least an order of magnitude higher than those actually observed. There must therefore be some ‘dirt’ on the surface that reduces the evaporation rate of the ices. Several possible hypotheses for mass loss from a ‘dirty’ comet follow:

  • The simplest is that mass loss is uniform over the cometary surface. This would result in a roughly uniform spherical cloud of dust and gas around the comet as the comet rotates. Comparing the predictions of such a model to pressure data measured by the COPS sensor in the ROSINA instrument suite on the Rosetta space craft suggests that this simple picture explains the data only moderately well (80% correlation coefficient).
  • A more sophisticated approach (but one that needs additional parameters) is that of localised activity, in which only a small fraction of the surface is active as ‘hot spots’. The numerical model is agnostic as to what the cause of such hot spots might be. This model does indeed give a better fit to the data, but this might be expected since it adds a parameter. However, such a model makes testable predictions about the velocity of dust particles entrained by the gas flows (higher than for uniform mass loss). A paper from the University of Bern group in Astronomy and Astrophysics (A&A 589 A90, 2016 – no open access) describes such a model and comparison to data.

Low-pressure gas flow and dust entrainment

When pressures are low enough, gas molecules no longer collide with each other very frequently, and the physics of gas flows changes from the regime we are familiar with (described by the Navier-Stokes equations for non-turbulent flows)  to that of rarefied gas dynamics which requires a different numerical approach. Which flow regime is relevant is determined by the Knudsen number, a ratio which essentially describes if a gas molecule is more likely to hit another gas molecule before it hits a surface or is lost. Such rarefied gas flows occur not only on and in comets, but also within vacuum systems and for rocket or engine exhausts in the upper atmosphere or space.

The numerical approach adopted by the MiARD project is that of Direct Simulation Monte Carlo. The experience of the University of Bern in using this technique for space science problems  is complemented by Heriot-Watt University’s experience in applying the technique to flow problems arising in the oil industry. The project will use this numerical approach to flows with a high Knudsen number in two different situations:

  1. For flows from the cometary surface into space
  2. For flows through the cometary soil to the surface, in particular to help set initial conditions for the flows from the comet (e.g. velocity distribution functions). For background, see e.g. “Predicting enhanced mass flow rates in gas microchannels using nonkinetic models” (not open access).

Further details from the University of Bern’s DRAG team.

What have we learnt from the Rosetta mission?

The shape and surface structure of 67P/Churyuomv-Gerasimenko were a big surprise. The comet is sometimes described as ‘duck shaped’, consisting of two lobes which probably have different densities, leading to speculation that this shape has a collisional origin. (Cracks at the neck of the comet between the lobes have been seen to increase during perihelion, and it could be that the comet will break into two parts eventually).

A future sample return mission?


Links to further information:


Ground-based observations

European Space Agency Rosetta site

Small comets FAQ

Astronomy Café FAQ for comets and meteorites

Miscellaneous space related info

International Association of Astronomical Artists

Planetary maps for children

Do NOT follow this link or you will be banned from the site!