Tag Archives: Rotation

The rotation of ‘Oumuamua

Hi there! Today we go back to ‘Oumuamua, you know, this interstellar object discovered last Fall. Its visit to our Solar system was the opportunity to observe it, and here we discuss on an analysis of the variations of its luminosity. I present you The excited spin state of 1I/2017 U1 ‘Oumuamua, by Michael J.S. Belton and collaborators. This study tells us that its rotation state might be complex, and that affects the way we figure out its shape. It has recently been published in The Astrophysical Journal Letters.

Remember 1I/’Oumuamua?

I already told you about ‘Oumuamua. This is the first identified object, which has been found in our Solar System but which undoubtedly originates from another System. In other words, it was formed around another star.
The Pan-STARRS survey identified ‘Oumuamua in October 2017, and the determination of its orbit proved it to be unusually eccentric. With an eccentricity close to 1.2, its orbit is a branch of a hyperbola rather than an ellipse. This means that it comes from very far, passes by while the Sun deviates it, and leaves us for ever.
This is the highest eccentricity ever recorded in the Solar System so far. Other objects had an eccentricity larger than 1, but which could have been caused by the gravitational perturbation of a planet. Not for ‘Oumuamua.
Its full name is actually 1I/2017 U1 (ʻOumuamua). 2017 because it was discovered in 2017, 1I as the first Interstellar object ever discovered (by the way, the International Astronomical Union has created this category for ‘Oumuamua), and the name ‘Oumuamua means scout in Hawaiian.

The announcement of its discovery motivated the observers all around the world to try to observe it and make photometric measurements. Here we discuss what these measurements tell us on the rotation and the shape. But before that, let me tell you something on the rotation.

Different modes of rotation

We will consider that our object is an ellipsoid. This is actually unsure, but let us assume it. We have 3 different axes, and we could imagine different configurations for its rotation:

  1. Tumbling rotation: the object rotates around its 3 axes, and basically this is a mess. We could be in a situation of dynamical chaos, like for the moon of Saturn Hyperion.
  2. Short-axis mode (SAM): the rotation is strongly dominated by a motion around the shortest axis. This is the case for many bodies in the Solar System, like the planets, our Moon… This does not mean that the rotation is strictly around one axis, but we will see that a little later.
  3. Long-axis mode (LAM): the rotation is strongly dominated by a motion around the longest axis.
The LAM and SAM modes.
The LAM and SAM modes.

These last two modes can actually cohabit with tumbling, i.e. a tumbling rotation may favor rotation around one axis.

If the rotation were strictly around one axis, then the body would look like a top. But this rotation axis may move with respect to the figure axis. This motion is named precession-nutation. The precession is the averaged path of the figure axis around the angular momentum, while the nutation contains the oscillations around it.

Now, imagine that you look at an object, which has such a rotation. How can you estimate it? There are ways.

Observing the rotation

Actually the brightness of a body not only depends on the distance from it, or on the insolation angle, but also on the surface facing you. This means that from the brightness, you can deduce something on the rotation state of the object. In particular, this surface brightness depends on its location with respect to the principal axis. If the object has the shape of a cigar, the reflected light from the long axis and from the short one will be different, and the lightcurve will present periodic variations. And the period of these variations is the rotation period. Easy, isn’t it?

Actually, not that easy. First, you assume that the surface has a constant albedo, i.e. that the ratio between the incident and the reflected lights is constant. But you do not know that. In particular, an icy surface has a higher albedo than a carbonaceous one. Another difficulty: a tumbling object, or even one with a precessional component in its rotation, will present a combination of different frequencies. Of course, this complicates the analysis.

However, you simplify the analysis in adding observations to your dataset. The authors used 818 observations over almost one month, spanning from Oct, 25 to Nov, 23, 2017. This includes observations from the Hubble Space Telescope, from the Magellan-Baade telescope at Las Campanas Observatory (Chile), from the Canada-France-Hawaii Telescope, from Pan-NSTARRS (these last facilities being based in Hawaii)…

Once the observations are obtained as raw data, they must be treated to correct from atmospheric and instrumental problems. And then it is not done yet, since the authors need an absolute luminosity of ‘Oumuamua, i.e. as if its distance to the observer were constant. The motion of ‘Oumuamua actually induced a trend in its distance to the Earth, and a trend in its luminosity, which the authors fitted before subtracting it the measured lightflux.

Once this is done, the authors get a lightcurve, which is constant on average, but presents variations around its mean value. Unfortunately, the required treatment induced an uncertainty in the measurements, which the authors had to consider. But fortunately, these practical difficulties are well-known, and algorithms exist to extract information from such data.

2 numerical algorithms

Basically, you need to extract periods from the variations of the lightflux. For that, we dispose of the classical tool of Fourier Transforms, which in principle requires equally spaced data. But the recorded data are not equally spaced, and remember that you must consider the uncertainties as well.

Specific algorithms exist for such a purpose. The authors used CLEAN and ANOVA, to double-check their results. These algorithms allow in particular to remove the aliasing effect, i.e. a wrong measurement of a period, because of an appropriate spacing of the data. And now, the results!

A cigar or a pancake?

The authors found two fundamental periods in the lightcurves, which are 8.67±0.34 and 3.74±0.11 hours. Interestingly, they connected these measurements to the possible dynamics of rotation, and they found two possible solutions:

  1. Long-Axis Mode: In that case, the possible rotation periods are 6.58, 13.15 and 54.48 hours, the latter being the most probable one.
  2. Short-Axis Mode: Here, ‘Oumuamua would be rotating with respect to the short-axis, but also with oscillations around the long axis of periods 13.15 or 54.48 hours.

In both axis, the long axis would also precess around the angular momentum in 8.67 ± 0.34 hours. Moreover, the authors found constraints on its shape. Previous studies already told us that ‘Oumuamua is highly elongated, this study confirms this fact, and tells us that ‘Oumuamua could be somewhere between the cigar and the pancake. But once more, this result could be weakened by variations of the surface albedo of ‘Oumuamua.

The study and its authors

And that’s it for today! Please do not forget to comment. You can also subscribe to the RSS feed, and follow me on Twitter, Facebook, Instagram, and Pinterest.

2010 JO179: a new, resonant dwarf planet

Hi there! Today I present you the discovery of a Trans-Neptunian Object, you know, these objects which orbit beyond the orbit of Neptune. And I particularly like that one, since its orbit resonates with the one of Neptune. Don’t worry, I will explain you all this, keep in mind for now that this object is probably one of the most stable. Anyway, this is the opportunity to present you A dwarf planet class object in the 21: 5 resonance with Neptune by M.J. Holman and collaborators. This study has recently been accepted for publication in The Astrophysical Journal Letters.

The Trans-Neptunian Objects

The Trans-Neptunians Objects are small bodies, which orbit beyond the orbit of Neptune, i.e. with a semimajor axis larger than 30 AU. The first discovered one is the well-known Pluto, in 1930. It was then, and until 2006, considered as the ninth planet of the Solar System. It was the only known TNO until 1992. While I am writing this, 2482 are listed on the JPL small-body database search engine.

The TNOs are often classified as the Kuiper-Belt objects, the scattered disc objects, and the Oort cloud. I do not feel these are official classifications, and there are sometimes inconsistencies between the different sources. Basically, the Kuiper-Belt objects are the ones, which orbits are not too much eccentric, not too inclined, and not too far (even if these objects orbit very far from us). The scattered disc objects have more eccentric and inclined orbits, and these dynamics could be due to chaotic / resonant excitation by the gravitational action of the planets. And the Oort cloud could be seen as the frontier of our Solar System. It is a theoretical cloud located between 50,000 and 200,000 Astronomical Units. Comets may originate from there. Its location makes it sensitive to the action of other stars, and to the Galactic tide, i.e. the deformation of our Galaxy.

The object I present you today, 2010 JO179, could be a scattered disc object. It has been discovered in 2010, thanks to the Pan-STARRS survey.

The Pan-STARRS survey

Pan-STARRS, for Panoramic Survey Telescope and Rapid Response System, is a systematic survey of the sky. Its facilities are located at Haleakala Observatory, Hawaii, and currently consist of two 1.8m-Ritchey–Chrétien telescopes. It operates since 2010, and discovered small Solar System objects, the interstellar visitor 1I/’Oumuamua… It observes in 5 wavelengths from infrared to visible.

The Pan-STARRS1 telescope. © Pan-STARRS
The Pan-STARRS1 telescope. © Pan-STARRS

The data consist of high-accuracy images of the sky, containing a huge amount of data. Beyond discoveries, these data permit an accurate astrometry of the object present on the images, which is useful for understanding their motion and determining their orbits. They also allow a determination of the activity of variable objects, i.e. variable stars, a study of their surface from their spectrum in the five wavelengths, and (for instance) the measurement of their rotation. A very nice tool anyway!

Pan-STARRS delivered its first data release in December 2016, while the DR2 (Data Release 2) is scheduled for mid-2018… pretty soon actually.

Among the discovered objects are the one we are interested in today, i.e. 2010 JO179.

Identifying the new object

The first observation of 2010 JO179 dates back from May 2010, and it has been detected 24 times during 12 nights, until July 2016. The detections are made in comparing the Pan-STARRS data from the known objects. Once something unknown appears in the data, leaving what the authors call a tracklet, its motion is extrapolated to predict its position at different dates, to investigate whether it is present on other images, another time. From 3 detections, the algorithm makes a more systematic search of additional tracklets, and in case of positive additional detection, then an orbit is fitted. The orbital characteristics (and other properties) are listed below.

Semimajor axis 78.307±0.009 AU
Eccentricity 0.49781±0.00005
Inclination 32.04342±0.00001 °
Orbital period 6663.757±0.002 yr
Diameter 600-900 km
Absolute magnitude 3.4±0.1

You can notice the high accuracy of the orbital parameters, which almost looks like a miracle for such a distant object. This is due to the number of detections, and the accuracy of the astrometry with Pan-STARRS. Once an object is discovered, you know where it is, or at least where it is supposed to be. Thanks to this knowledge, it was possible to detect 2010 JO179 on data from the Sloan Digital Sky Survey, taken in New Mexico, and on data from the DECalS survey, taken in Chile. Moreover, 2010 JO179 was intentionally observed with the New Technology Telescope (NTT) in La Silla, Chile.

The spectroscopy (analysis of the reflected light at different wavelengths) of 2010 JO179 revealed a moderately red object, which is common for TNOs.

Measuring its rotation

This is something I have already evoked in previous articles. When you record the light flux reflected by the surface of a planetary body, you should observe some periodic variability, which is linked to its rotation. From the observations, you should extract (or try to) a period, which may not be an easy task regarding the sparsity and the accuracy of the observations.

In using the so-called Lomb-Scargle algorithm, the authors detected two possible periods, which are 30.6324 hours, and 61.2649 hours… i.e. twice the former number. These are best-fits, i.e. you try to fit a sinusoid to the recorded light, and these are the periods you get. The associated amplitudes are variations of magnitude of 0.46 and 0.5, respectively. In other words, the authors have two solutions, they favor the first one since it would imply a too elongated asteroid. Anyway, you can say that twice 30.6324 hours is a period as well, but what we call the spin period is the smallest non-null duration, which leaves the light flux (pretty) invariant. So, the measured spin period of 2010 JO179 is 30.6324 hours, which makes it a slow rotator.

Mean-motion resonances

Let us make a break on the specific case of 2010 JO179 (shall we give it a nickname anyway?), since I would like to recall you something on the mean-motion resonances before.

When two planetary bodies orbit the Sun, they perturb each other. It can be shown that when the ratio of their orbital periods (similarly the ratio of their orbital frequencies) is rational, i.e. is one integer divided by another one, then you are in a dynamical situation of commensurability, or quasi-resonance. A well known case is the 5:2 configuration between Jupiter and Saturn, i.e. Jupiter makes 5 orbits around the Sun while Saturn makes 2. In such a case, the orbital perturbations are enhanced, and you can either be very stable, or have a chaotic orbit, in which the eccentricities and inclinations could raise, the orbit become unpredictable beyond a certain time horizon (Lyapunov time), and even a body be ejected.

Mathematically, an expansion of the so-called perturbing function, or the perturbing mutual gravitational potential, would display a sum of sinusoidal term containing resonant arguments, which would have long-term effects. These arguments would read as pλ1-(p+q)λ2+q1ϖ1+q2ϖ2+q3Ω1+q4Ω2, with q=q1+q2+q3+q4. The subscripts 1 and 2 are for the two bodies (in our case, 1 will stand for Neptune, and 2 for JO 2010179), λ are their mean longitudes, ϖ their longitudes of pericentres, and Ω the longitudes of their ascending nodes.

In a perturbed case, which may happen for high eccentricities and inclinations, resonances involving several arguments may overlap, and induce a chaotic dynamics that could be stable… or not. You need to simulate the long-term dynamics to know more about that.

A resonant long-term dynamics

It appears that Neptune and 2010 JO179 are very close to the 21:5 mean-motion resonance (p=5, q=16). To inquire this, the authors ran 100 numerical simulations of the orbital motion of 2010 JO179, with slightly different initial conditions which are consistent with the uncertainty of the observations, over 700 Myr. And they saw that 2010 JO179 could be trapped in a resonance, with argument 5λ1-21λ2+16ϖ2. In about 25% of the simulations, JO179 remains trapped, which implies that the resonant argument is librating, i.e. bounded, all over the simulation. As a consequence, this suggests that its orbit is very stable, which is remarkable given its very high eccentricity (almost 0.5).

The study and its authors

And that’s it for today! Please do not forget to comment. You can also subscribe to the RSS feed, and follow me on Twitter, Facebook, Instagram, and Pinterest.

Rough terrains spin up asteroids

Hi there! If you follow me, you have already heard of the Yarkovsky effect, or even of the YORP, which are non-gravitational forces affecting the dynamics of Near-Earth Asteroids. Today I tell you about the TYORP, i.e. the Tangential YORP. This is the opportunity for me to present you Analytic model for Tangential YORP, by Oleksiy Golubov. This study has recently been published in The Astronomical Journal. The author meets the challenge to derive an analytical formula for the thermal pressure acting on the irregular regolith of an asteroid. Doing it requires to master the physics and make some sound approximations, following him tells us many things on the Tangential YORP.

From Yarkovsky to TYORP

When we address the dynamics of Near-Earth Asteroids, we must consider the proximity of the Sun. This proximity involves thermal effects, which significantly affect the dynamics of such small bodies. In other words, the dynamics is not ruled by the gravitation only. The main effect is the Yarkovsky effects, and its derivatives.


The Sun heats the surface of the asteroid which faces it. When this surface element does not face the Sun anymore, because of the rotation of the asteroid, it cools, and radiates some energy. This effect translates into a secular drift in the orbit, which is known as the Yarkovsky effect. This Yarkovsky effect has been directly measured for some asteroids, in comparing the simulated orbit from a purely gravitational simulation, with the astrometric observations of the objects. Moreover, long-term studies have shown that the Yarkovsky effect explains the spreading of some dynamical families, i.e. asteroids originating from a single progenitor. In that sense, observing their current locations proves the reality of the Yarkovsky effect.
When the asteroid has an irregular shape, which is common, the thermal effect affects the rotation as well.


Cooling a surface element which has been previously heated by the Sun involves a loss of energy, which depends on the surface itself. This loss of energy affects the rotational dynamics, which is also affected by the heating of some surface. But for an irregular shaped body, the loss and gain of energy does not exactly balance, and the result is an asteroid which spins up, like a windmill. In some cases, it can even fission the body (see here). This effect is called YORP, for Yarkovsky-O’Keefe–Radzievskii–Paddack.

This is a large-scale effect, in the sense that it depends on the shape of the asteroid as a whole. Actually, the surface of an asteroid is regolith, it can have boulders… i.e. high-frequency irregularities, which thus will be heated differently, and contribute to YORP… This contribution is known as Tangential YORP, or TYORP.

Modeling the physics

When you heat a boulder from the Sun, you create an inhomogeneous elevation of temperature, which can be modeled numerically, with finite elements. For an analytical treatment, you cannot be that accurate. This drove the author to split the boulder into two sides, the eastern and the western sides, both being assumed to have an homogeneous temperature. Hence, two temperatures for the boulder. Then the author wrote down a heat conduction equation, which says that the total heat energy increase in a given volume is equal to the sum of the heat conduction into this volume, the direct solar heat absorbed by its open surface, and the negative heat emitted by the open surface (which radiates).

These numbers depend on

  • the heat capacity of the asteroid,
  • its density,
  • its heat conductivity,
  • its albedo, i.e. its capacity to reflect the incident Solar light,
  • its emissivity, which characterizes the radiated energy,
  • the incident Solar light,
  • the time.

The time is critical since a surface will heat as long it is exposed to the Sun. In the calculations, it involves the spin frequency. After manipulation of these equations, the author obtains an analytical formula for the TYORP pressure, which depends on these parameters.

A perturbative treatment

In the process of solving the equations, the author wrote the eastern and western temperatures as sums of periodic sinusoidal solutions. The basic assumption, which seems to make sense, is that these two quantities are periodic, the period being the rotation period, P, of the asteroid. This implicitly assumes that the asteroid rotates around only one axis, which is a reasonable assumption for a general treatment of the problem.
As a result, the author expects these two temperatures to be the sum of sines and cosines of periods P/n, P being an integer. For n=1, you have a variation of period P, i.e. a diurnal variation. For n = 2, you have a semi-diurnal one, etc.

The perturbative treatment of the problem consists in improving the solution in iterating it, first in expressing only one term, i.e. the diurnal one, then in using the result to derive the second term, etc. This assumes that these different terms have amplitudes, which efficiently converge to 0, i.e. the semi-diurnal effect is supposed to be negligible with respect to the diurnal one, but very large with respect to the third-diurnal, etc. Writing down the solution under such a form is called Fourier decomposition.

The author says honestly that he did not check this convergence while solving the equation. However, he successfully tested the validity of his obtained solution, which means that the resolution method is appropriate.


The author is active since many years on the (T)YORP issue, and has modeled it numerically in a recent past. So, validating his analytical formula consisted in confronting it with his numerical results.

He particularly confronted the two results in the cases of a wall, a half buried spherical boulder, and a wave in the regolith, with respect to physical characteristics of the material, i.e. dimension and thermic properties. Even though visible differences, the approximation is pretty good, validating the methodology.

This allowed then the author to derive an analytical formula of the TYORP pressure on a while regolith, which is composed of boulders, which sizes are distributed following a power law.


This is the first analytical formula for the TYORP, and I am impressed by the author’s achievement. We can expect in the future that this law (should we call it the Golubov law?) would be a reference to characterize the thermic properties of an asteroid. In other words, future measurements of the TYORP effect could give the thermic properties, thanks to this law. This is just a possibility, which depends on the reception of this study by the scientific community, and on future studies as well.

The study and its author

And that’s it for today! Please do not forget to comment. You can also subscribe to the RSS feed, and follow me on Twitter and Facebook.

Rotation and activity of a comet

Hi there! We, Earthians, are regularly visited by periodic comets, the most famous one being probably 1P/Halley, which will visit us in 2061. Since we cannot wait, we study others of that kind. Today I tell you about 49P / Arend-Rigaux. This is the opportunity for me to present you The rotation and other properties of Comet 49P/Arend-Rigaux, 1984 – 2012, by Nora Eisner, Matthew M. Knight and David G. Schleicher. This study has recently been published in The Astronomical Journal.

The comet 49P / Arend-Rigaux

The comet 49P / Arend-Rigaux has been discovered in February 1951 at the Royal Observatory of Belgium, by Sylvain Arend and Fernand Rigaux. It is a periodic comet of the Jupiter family, i.e. with a period smaller than 20 years. Its period is actually 6.71 years, its semimajor axis 3.55 AU (astronomical units, 1 AU being 150 millions km, i.e. the Sun-Earth distance), its eccentricity 0.6, and its orbital inclination 19°, with respect to the ecliptic. These numbers are extracted from the JPL Small-Body Database Browser, and are calculated at the date Apr 6, 2010. I have plotted below the distances Sun-comet and Earth-comet.

Distance to the Sun.
Distance to the Sun.
Distance to the Earth.
Distance to the Earth.

The distance to the Sun clearly shows the periodic variations. The orbit of the Earth is at 1 AU, the one of Mars at 1.5 AU, and the one of Jupiter at 5.2 AU. Every 6.71 years, the comet reaches its perihelion, i.e. minimizes its distance to the Sun. This proximity warms the comet and provokes an excess of cometary activity, i.e. sublimation of dirty ice. At these occasions, the distance with the Earth is minimized, but with variations due to the orbital motion of the Earth. We can see for instance that the comet gets pretty close to the Earth in 1951 (when it was discovered), in 1984, and in early 2032. These are favorable moments to observe it. The paper I present you today is mainly (but not only) based on photometric observations made between January and May 2012, at Lowell Observatory.

Observations at Lowell Observatory

Lowell Observatory is located close to Flagstaff, AZ (USA). It was founded by the famous Percival Lowell in 1894, and is the place where Clyde Tombaugh discovered Pluto, in 1930. Among its facilities is the 4.28 m Discovery Channel Telescope, but most of the data used in this study were acquired with the 1.1 m Hall telescope, which is devoted to the study of comets, asteroids, and Sun-like stars. The authors also used a 79 cm telescope. The observations were made in the R(ed) band.

The data

Besides these 33 observation nights during the first half of 2012, the authors used data acquired close to the 1984 and 2005 perihelion passages, even if the 2005 ones revealed unusable. The observations consists to measure the magnitude (somehow, the luminosity) of the comet, in correcting for atmospheric problems, so as to be able to detect the variations of this magnitude. You can find below an example of data:

Magnitude of 49P / Arend-Rigaux measured in April 2012.
Magnitude of 49P / Arend-Rigaux measured in April 2012.

Of course, the data have holes, since you cannot observe during the day. Moreover, the comet needs to be visible from Arizona, otherwise it was just impossible to observe it and make any measurements.

We can see a kind of periodicity in the magnitude, this is a signature of the rotation of the comet.

Measuring the rotation

Most of the planetary bodies are kinds of triaxial ellipsoids. Imagine we are in the equatorial plane of one of them. We see an alternation of the long and short axes of its equatorial section. If the albedo of the surface element we face depends mainly on its curvature (it depends on it, but mainly may be an overstatement), then we should see two peaks during a period. As a consequence, the period of the lightcurve we observe should be half the rotation period of the comet.

In combining all the measurements, the authors managed to derive a rotation period of 13.45 ± 0.01 hour. For that, they used two different algorithms, which gave very close results, giving the authors confidence in their conclusions. The first one, Phase Dispersion Minimization (PDM), consists to assume a given period, split the measurements into time intervals of this period, and overlap them. The resulting period gives to the best overlap. The other algorithm is named Lomb-Scargle, following its authors. It is a kind of Discrete Fourier Transform, but with the advantage of not requiring uniformly sampled data.

In addition to this rotation period, the authors detected an increasing trend in the 2012 data, as if the spin of the comet accelerated. This is in agreement with an alteration of the measured rotation from the Earth, which moves, and reveals a retrograde rotation, i.e. an obliquity close to 180°. In other words, this is an illusion due to the motion of the observer, but this illusion reveals the obliquity.

Moreover, in comparing the 2012 data with the ones of 1984, the authors managed to detect a variation in the rotation period, not larger than 54 seconds. This is possible regarding the fact that the comet is altered by each perihelion passage, since it outgasses. In this case, that would imply a change of at the most 14 seconds of the rotation period between two passages. Such variations have also been detected for at least 4 other comets (2P/Encke, 9P/Tempel 1, 10P/Tempel 2, and 103P/Hartley 2, see Samarinha and Mueller (2013)).

Comet Period (h) Variation (s)
2P/Encke 11 240
9P/Tempel 1 41 -840
10P/Tempel 2 9 16.2
103P/Hartley 2 18 7200
49P/Arend-Rigaux 13.45 -(>14)

Finally, since the lightcurve is a signature of the shape as well, the authors deduced from the amplitude of variation that the axial ratio of the nucleus, i.e. long axis / short axis, should be between 1.38 and 1.63, while an independent, previous study found 1.6.

Cometary activity

49P / Arend-Rigaux has a low activity. Anyway, the authors detected an event of impulse-type outburst, which lasted less than 2 hours. The analysis of the coma revealed an excess of cyanides with respect to the 1984 passage. Moreover, 49P / Arend-Rigaux is the first comet to show hydroxyde.

The study and its authors

That’s it for today! Please do not forget to comment. You can also subscribe to the RSS feed, and follow me on Twitter and Facebook.

Indirect measurement of an asteroid’s pole

Hi there! Today, another paper on the Yarkovsky effect. You know, this non-gravitational force which acts on the asteroid, especially if it is close enough to the Sun. After reading this post, you will know how it can reveal us the obliquity of an asteroid. I present you Constraints on the near-Earth asteroid obliquity distribution from the Yarkovsky effect, by C. Tardioli, D. Farnocchia, B. Rozitis, D. Cotto-Figueiroa, S.R. Chesley, T.S. Statler & M. Vasile. This paper has recently been accepted for publication in Astronomy and Astronomy.

The way it works

Imagine you want to know the rotation of an asteroid… but you cannot measure it directly. However, you can measure the orbital motion of the asteroid, with enough accuracy to detect an effect (here Yarkovsky), which itself depends on the rotation… measuring Yarkovsky is measuring the rotation! Easy, isn’t it?

The rotation of an asteroid

As any planetary body, an asteroid has a rotational motion, which consists in spinning around one axis (actually 3, but you can safely neglect this fact), at a given rate. We can consider that we know its rotation when

  1. We know its spin rate, or its rotational period (let us assume it is constant),
  2. We know the orientation of its spin pole. We will call it the obliquity.

Usually the asteroids spin in a few hours, which is very fast since they need at least several months to complete one revolution around the Sun. The obliquity is between 0° and 180°. 0° means that the spin axis is orthogonal to the orbital plane, and that the rotation is prograde. However, 180° is the other extreme case, the spin axis is orthogonal, but with a retrograde rotation.

A direct measurement of these two quantities would consist in following the surface of the asteroid, to observe the rotation. Usually we cannot observe the surface, but sometimes we can measure the variations of the magnitude of the asteroid over time. This is directly due to the Solar light flux, which is reflected by the surface of the asteroid. Because the topography is irregular, the rotation of the asteroid induces variations of this reflection, and by analyzing the resulting lightcurve we can retrieve the rotational quantities.

Very well, but sometimes the photometric observations are not accurate enough to get these quantities. And other times, the measured rotational quantities present an ambiguity, i.e. 2 solutions, which would need an independent measurement to discriminate them, i.e. determine which of the two possible results is the right one.

It appears that the Yarkovsky effect, which is an alteration of the orbital motion of the body due to the inhomogeneity of its temperature, itself due to the Solar incident flux and the orientation of the body, i.e. its rotation, can sometimes be measured. When you know Yarkovsky, you know the obliquity. Well, it is a little more complicated than that.

Yarkovsky: A thermal effect

Since I have already presented you Yarkovsky with words, I give you now a formula.

The Yarkovsky effect, i.e. the thermal heating of the asteroid, induced a non-gravitational acceleration of its orbital motion. This acceleration reads A2/r2, where r is the distance to the Sun (remember that the asteroid orbits the Sun), and

A2 = 4/9(1-A)Φ(αf(θs)cos(ε)-f(θo)sin2(ε)),


  • A: albedo of the asteroid, i.e. quantity of the reflected light wrt the incident one,
  • Φ: Solar radiation,
  • α: an enhancement factor. This is a parameter…
  • ε: the obliquity (which the authors determined),
  • θs / θo: thermal parameters which depend on the spin period (s), and the orbital one (o), respectively.

If you know Yarkovsky, you know A2, since you know the distance r (you actually know where the asteroid is). If you know all the parameters except ε, then A2 gives you ε. In fact, some of the other parameters need to be estimated.

Measuring Yarkovsky

As you can see, this study is possible only for asteroids, for which you can know the Yarkovsky acceleration. Since it is a thermal effect, you can do it only for Near-Earth Asteroids, which are closer to the Sun than the Main Belt. And to measure Yarkovsky, you must simulate the orbital motion of the asteroid, which is perturbed by the main planets and Yarkovsky, with the Yarkovsky acceleration as a free parameter. A fit of the simulations to the actual astrometric observations of the asteroid gives you a number for the Yarkovsky acceleration, with a numerical uncertainty. If your number is larger than the uncertainty, then you have detected Yarkovsky. And this uncertainty mainly depends on the accuracy of your astrometric observations. It could also depend on the validity of the dynamical model, i.e. on the consideration of the forces perturbing the orbital motion, but usually the dynamical model is very accurate, since the masses and motions of the disturbing planets are very well known.
The first detection of the Yarkovsky acceleration was in 2003, when a drift of 15 km over 12 years was announced for the asteroid 6489 Golevka.

So, you have now a list of asteroids, with their Yarkovsky accelerations. The authors worked with a final dataset of 125 asteroids.

So many retrograde asteroids

The authors tried to fit a distribution of the obliquities of these asteroids. The best fit, i.e. which reduces the distance between the resulting obliquities and the Yarkovsky acceleration that they would have produced, is obtained from a quadratic model, i.e. 1.12 cos2(ε)-0.32 cos(ε)+0.13, which is represented below.

Distribution of the asteroids with respect to their obliquity.
Distribution of the asteroids with respect to their obliquity.

What you see is the number of asteroids with respect to their obliquity. The 2 maxima at 0° and 180° mean that most of the asteroids spin about an axis, which is almost orthogonal to their orbital plane. From their relative heights, it appears that there about twice more retrograde asteroids than prograde ones. This is consistent with previous studies, these obliquities actually being a consequence of the YORP effect, which is the influence of Yarkovsky on the rotation.

The study and its authors

That’s it for today! Please do not forget to comment. You can also subscribe to the RSS feed, and follow me on Twitter and Facebook.