Category Archives: Phenomena

The Solar System’s Place In The Milky Way: Part 1

You are here in the Milky Way galaxy
Click to enlarge (with local arms labelled)
Base map: NASA / JPL-Caltech / R. Hurt (SSC-Caltech)

When I wrote recently about the pole stars of other planets, I was aware of one thing my sky maps didn’t show—the rotation poles of our galaxy. They weren’t really relevant to that discussion, but I’m now prompted to write a bit about the Milky Way galaxy, and our relationship to it, because I’ve just encountered a rather garbled, misleading and self-contradictory Wikipedia article on the topic—specifically the section entitled Sun’s location and neighborhood. (The last time I referred to a misleading Wikipedia page on this blog, the page was eventually fixed by an editor citing my article. This was gratifying, but it meant my link to the page no longer demonstrated the problem to which I’d objected. So this time I’ve linked to a Wayback Machine copy of the page dated 9 April 2023, which will at least give permanent context to my griping.)

The view of our galaxy at the head of this post is an artist’s impression, sketching in the major features as currently understood. Our sun, marked by the arrow, lies between two spiral arms. The Perseus Arm, which lies farther out than our location, is considered to be one of the Milky Way’s two major spiral arms; and inwards from our position is the more minor Sagittarius Arm. Spanning the gap between these two arms is a more diagonal structure, clearly visible in the painting, variously referred to as the Orion Spur, the Orion Arm, or the Local Arm, the status of which seems to be much debated. Our sun lies within the inner edge of that diagonal structure.*

Our distance from the centre of the galaxy isn’t known with great accuracy. From a couple of recent papers addressing the issue, the distance is somewhere in the region of 7.9 kiloparsecs (25,800 light-years) to 8.34 kiloparsecs (27,200 light-years), which you’ll see from the map above is very roughly halfway from the galactic centre to its rim. The two papers in my link are in much stronger agreement about the angular velocity of the sun in its orbit around the galaxy, which they place between 30.2 and 30.6 km/s/kpc. Call it 30.4, which translates to 9.85×10-16rad/s, or one orbit every 200 million years. But the sun is near perigalacticon at present (its closest approach to the galactic centre) and therefore moving faster than it normally does. Commonly quoted estimates for the duration of a true galactic year therefore lie between 225 and 250 million years.

So much for our radial location within the galaxy. How are we placed relative to the galactic plane? Estimates of our distance from the galactic mid-plane (which we can think of as the galactic equator) vary, but a recent measurement and review places it about 17 parsecs (55 light-years) “above” the plane. I’ll come back to what “above” means in this context later.

The orientation of the solar system relative to the plane of the galaxy turns out to be a little unexpected. In the view below, we’ve moved a short distance from the sun towards the galactic rim, and are looking back on the solar system framed against the nebulae of the galactic core, with the galactic equator running horizontally across the picture.

Orientation of the plane of the solar system relative to the galaxy
Click to enlarge
Prepared using Celestia and the “Milky Way” add-on by Guillermo Abramson

I’ve marked the constellations in the vicinity of the galactic core for orientation. The four visible orbits are those of the outer giant planets from Jupiter to Neptune.

So the plane of the solar system is tilted at an angle of about 60° relative to the plane of the galaxy. It also, as you see, comes close to aligning with the centre of the galaxy—the mismatch is only about 6°. Since the plane of the solar system remains fixed as it orbits the galaxy, a perfect alignment will occur twice in every galactic year—the most recent happened three or four million years ago.

Now I’m going to zoom in to look at the Earth’s orbit. The plane of the Earth’s orbit is called the ecliptic, and it’s pretty closely aligned with the overall plane of the solar system—close enough that I plan on using it as a reasonable proxy for the solar system’s invariable plane, later.

Orientation of Earth's orbit relative to the galaxy
Click to enlarge
Prepared using Celestia and the “Milky Way” add-on by Guillermo Abramson

I’ve included a cartoonishly large Earth to give you a visual cue to the orientation of the ecliptic along the axis perpendicular to your screen—the left side of the Earth’s orbit, as seen here, is farther away from our viewpoint than is the sun, while the right side brings the Earth closer to us. I’ve also marked four key locations in the Earth’s orbit—the solstices and equinoxes. And here’s a coincidence! The solstice points align pretty well with the galactic core. In 2023, for example, the sun will cross the galactic equator at about 11:00 on 22 December, the day of the solstice.

Now, we know that the Earth’s axis is tilted by about 23½° relative to the ecliptic. So how does the Earth’s tilt interact with the solar system’s tilt? There’s a hint in my previous picture, from the location of the solstices and equinoxes, but we can zoom in farther from the same vantage point to make the situation clearer.

Orientation of the Earth relative to the galaxy
Click to enlarge
Prepared using Celestia and the “Milky Way” add-on by Guillermo Abramson

The Earth is rotated in the plane of the image through the 60-degree tilt of the rest of the solar system, but the 23½°-tilt between its equator and its orbit is directed towards us, out of the computer screen. In other words, the Earth’s tilt relative to the ecliptic is almost at right angles to the ecliptic’s tilt relative to the galaxy. This is difficult to depict in a single diagram, like the one offered by Astronomy magazine recently (third image down on the linked page), in which it looks very much as if the Earth’s tilt is in the same plane as the ecliptic tilt. And it’s something that can catch out even the astronomers who wrote the entries for Swinburne University’s online Encyclopedia of Astronomy. For a few years their entry on the Galactic Plane claimed that:

The galactic plane is tilted at an angle of 63 degrees to the celestial equator. Since the ecliptic (the path of the Sun on the sky) is inclined at an angle of 23.5 degrees to the celestial equator, the galactic plane and the ecliptic are nearly at right angles (63 + 23.5 = 86.5 degrees), although this is purely coincidental.

Purely fantastical, more like. The entry was mercifully corrected in 2012; my link takes you to the original version embarrassingly preserved on the Wayback Machine.

But they were correct about the 63° tilt between the celestial equator (the extension of the Earth’s equator into the sky) and the galactic plane. The easiest way to see how all this fits together is not to try to diagram the intersecting planes of the Earth’s equator, the ecliptic and the galactic equator, but instead to plot the position of their north poles in the sky. The Earth has its celestial north pole, near the pole star, Polaris. The ecliptic has its own north pole, corresponding to the rotation axis implied by the movement of the planets in their orbits. And galactic north was defined by the International Astronomical Union, back in 1958, as being the extension of the rotation axis of the galaxy into the northern sky of the Earth. (Which, at last, allows me to say what is meant by the sun sitting about 55 light-years “above” the galactic plane—it lies north of the galactic plane.)

Position of celestial, ecliptic and galactic north poles
Click to enlarge

(As with my previous post about the pole stars of other planets, I’ve used a star map generated by In-the-sky.org.)

The north galactic pole lies in the obscure little constellation Coma Berenices. And you can immediately see how the angle between the ecliptic and celestial north poles is measured almost at right angles to the angle between the galactic and ecliptic north poles, so that the celestial and ecliptic poles end up at much the same angular distance from the galactic pole.

But there’s one little quirk to this trio of north poles, arising from the way the IAU defines north, which I addressed in more detail when I wrote about pole stars previously. If we looked down on Earth from the celestial north pole, we’d see it rotating anticlockwise; likewise, if we looked at the solar system from the ecliptic north pole, we’d see the planets moving anticlockwise around the sun. But if we were to look at the galaxy from the galactic north pole (the view in the picture at the head of this post), we’d observe it rotating clockwise. Like the planet Venus, then, the Milky Way galaxy turns out to be a retrograde rotator.

I can now go back to my original solar system diagram, and add some orientation arrows.

Solar System Orientation
Click to enlarge
Prepared using Celestia and the “Milky Way” add-on by Guillermo Abramson

The solar system, which participates in the galactic rotation, is therefore moving in a direction approximately aligned with my horizontal arrow. The extent to which it deviates from that alignment will be my topic next time I write on this topic—at which point I’ll also be able to explain my gripes about the Wikipedia article that started all this.


* For more detail on the layout of the spiral arms, and detailed maps of the solar neighbourhood, visit Kevin Jardine’s jaw-dropping Galaxy Map site, which is a thing of beauty and a labour of love.
This approximate alignment between the galactic core, the sun and the Earth on the day of the December solstice was of course one of the foundations of a fatuous end-of-the-world scenario predicted for 2012. (To be honest, though, I did quite enjoy Roland Emmerich’s accompanying disaster movie, 2012).
There’s a passage in Larry Niven’s classic science fiction novel, Ringworld (1970), in which he describes how the Puppeteer Fleet of Worlds abandons the Milky Way galaxy and sets out for the Small Magellanic Cloud. Elsewhere in the novel, Niven tells us that the Fleet of Worlds is “moving north along the galactic axis”. Unfortunately, the SMC actually lies south of the galactic equator. Oops. But this would make sense if Niven was using rotational north (that is, the direction from which a body appears to rotate anticlockwise), rather than the IAU’s parochial system based on the orientation of the Earth. (I once actually raised this possibility with Niven, but he responded that the Puppeteers were “taking the scenic route”.)

Converting Apollo State Vectors To Orbits

Apollo 11 Translunar trajectory relative to Van Allen Belts 3
Click to enlarge
Prepared using Celestia

OK, another mathematical one. This is the fourth in a series of posts about the orbits followed by the Apollo spacecraft as they travelled to and from the moon—something I suppose is getting a little more topical now that NASA has finally got underway with its planned return to the lunar surface.

I started with a post entitled “Keplerian Orbital Elements”, which introduced the various parameters used to describe an orbit—these are the numbers you need to plug into a piece of orbit-plotting software, like Celestia, so that it will display the spacecraft’s trajectory for you. (It’s what I used to prepare the diagram at the head of this post.)

Then I progressed to “Finding Apollo Trajectory Data”, in which I provided links to the original Apollo documentation, and described how to pull the necessary data from those sources.

Then I digressed into “The Advent Of Atomic Time”, as a way of introducing the difference between the GMT times listed in the Apollo trajectory documents, and the Terrestrial Time (TT) we need to use in order to correctly describe the Apollo orbits.

In the first post I established that we need to know the following six orbital elements:

  • semimajor axis (a)
  • eccentricity (e)
  • inclination (i)
  • longitude of the ascending node (Ω)
  • argument of the periapsis (ω)
  • true anomaly (θ)

The first two give the orbit’s size and shape, the next three its orientation in three dimensions, and the last one tells us the spacecraft’s position in orbit—see my first post for more detail.

To go along with the true anomaly we need the time, called the epoch (t), at which our spacecraft is in that position. This is usually expressed in the form of a Julian Day, and there are online calculators that will convert a date and Greenwich Mean Time to a Julian Day—there’s one here.

And we need an orbital period (P). We can calculate this from the semimajor axis, if we know the mass (M) of the central body around which the orbiting spacecraft moves—in this case, the Earth. There’s a conversion factor, too, called the Universal Gravitational Constant (G), but since orbital calculations always involve multiplying M by G, and since M is constant for any given planet, the two can be lumped together into something called the Standard Gravitational Parameter (μ) for that planet. Currently accepted values for μ for the various planets are listed by the Jet Propulsion Laboratory.

In my second post I established that what we have available for Apollo are primarily state vectors—three-dimensional locations and velocities for given times. Here, for instance, is the Apollo 11 state vector for the time of Translunar Injection (TLI), which established the initial orbit on which the Apollo spacecraft departed from Earth.

TLI state vector for Apollo 11
Click to enlarge (Source: Apollo/Saturn V Postflight Trajectory AS-506)

I’ve marked the relevant data columns in red, and we have enough information here to produce all the required orbital elements and an epoch. We have, from left to right:

  • ground elapsed time (GET)
  • geocentric distance (r)
  • longitude (λ)
  • geocentric latitude (ψ)
  • heading (h)
  • flight-path angle (f)
  • space-fixed velocity (v)

I explained all these in my second post, and I’ll refer you back to that for more detail. GET measures the elapsed time since launch, and by adding that to the Julian Day calculated for the launch date and time, we get our epoch (t) for TLI—I walked through that process step by step in my second post. The next three parameters describe the spacecraft’s position in three-dimensional space, but relative to the rotating surface of the Earth, so there’s a little bit of work to do to convert these coordinates to the fixed celestial coordinate system we need for our orbit. And the next three numbers give the spacecraft’s velocity, again in three dimensions, and NASA have helpfully done the calculation to make this “space-fixed”—that is, they’ve taken the spacecraft’s velocity relative to the surface of the Earth (given in the “EF VEL” column), and added in the rotation speed of the Earth at that latitude, so that the quoted velocity is now relative to the fixed stars, rather than the rotating Earth.

In my second post, I worked out the epoch (in Julian Days) of Apollo 11’s Translunar Injection:

tUTC = 2440419.18209525 days

But, as I explained in my third post, this time is in Coordinated Universal Time (UTC), and to that figure we need to add the difference between UTC and TT on 16 July 1969, the date of the Apollo 11 Translunar Injection. In that previous post, I established that this difference, symbolized ΔT, was 39.746 seconds, and I’ll refer you back to that post for a description of the relevance and origin of that number.

So the Terrestrial Time epoch for our Apollo 11 orbit is:

tTT = tUTC +ΔT

tTT = 2440419.18255527 days

Now I need to convert the latitude and longitude coordinates to the equatorial coordinate system used in astronomy. The equivalent of latitude in this system is called declination (δ), and longitude translates into right ascension (α). The good news is that declination is equal to the geocentric latitude, so we can just set δ equal to ψ:

δ = 9.9204°

The conversion between λ and α is complicated by the Earth’s rotation. The zero meridian of longitude (the Greenwich meridian) only points at the zero meridian of right ascension (the vernal equinox) once per rotation. So to work out the right ascension corresponding to the longitude of TLI, we need to know in which direction the Greenwich meridian was pointing at time t, and this is a little complicated by the fact that the Earth doesn’t rotate at a constant rate. The angle between the vernal equinox and the Greenwich meridian is given by a quantity called Greenwich Mean Sidereal Time (GMST), which I introduced in my third post. Although this is commonly expressed in hours, minutes and seconds, it can be converted to an angular measurement in degrees, which is what we need. To find GMST we ideally need to know Universal Time in the form of UT1, which is approximated by the civil time, UTC. It was actually quite well approximated by UTC during the time of the earlier Apollo missions, and I looked up the difference between the two times on 16 July 1969 in my third post:

ΔUT1 = 0.0115 seconds

So to calculate GMST properly we need another version of the epoch:

tUT1 = tUTC + ΔUT1

tUT1 = 2440419.18209538 days

You’ll find all sorts of equations to convert between UT1 and GMST. The one I use below comes from Chapter 2 of a Jet Propulsion Laboratory publication entitled Explanatory Supplement to Metric Prediction Generation—see equation 2-15, but the whole chapter is a nice primer on the complexities of astronomical time measurements.

First we need to express our UT1 epoch in terms of Julian Days elapsed since midday on 1 January 2000:

d = tUT12451545 days

Then we plug this value d into a rather large equation it’s difficult to cram on to one line:

\tiny GMST = 280.460618375 + 360.9856473663\cdot d + 2.9079\times 10^{-13}\cdot d^{2}-5.3\times 10^{-22}\cdot d^{3}

This usually generates a big number, reflecting the many rotations that have taken place between 1 January 2000 and the epoch of interest, so we need to narrow it down to a value between 0 and 360 degrees by adding or subtracting whole multiples of 360. Most spreadsheets and programming languages let you do this using the modulo function, which gives the remainder after division. So mod(GMST,360), or something similar, will give the remainder after GMST is divided by 360, which is what we need.*

When I run the TLI epoch through this process, I get an answer of -180.1181 degrees—negative, because 1969 comes before the year 2000. This is the equivalent of 179.8819 degrees, so at the epoch of TLI, the Greenwich meridian was pointing towards a right ascension of 179.8819 degrees.

Our longitude (λ) from the NASA table is -164.8373, and the right ascension is just λ+GMST, giving us, at long last:

α = 15.0446°

So that’s the fiddly bit done, and we now have our state vector in the necessary space-fixed coordinates.

Many reference sources deal with the conversion from state vector to orbital parameters. Ulrich Walter’s magisterial Astronautics: The Physics of Space Flight is a great resource, but uses vector notation I’m going to avoid here. A useful on-line resource is Uger Guven’s slide collection, Satellite Orbit Dynamics.

For a given state vector, three of the orbital parameters depend on the planetary mass. To see why this is so, consider a spacecraft with exactly the same position and velocity as Apollo 11’s at TLI, but orbiting a planet with half the mass—it will go farther from the planet before falling back again under the influence of gravity. So it’s useful to define a constant, C, which I’ll call the “energy constant”, because it’s a measure of the ratio between the spacecraft’s gravitational potential energy and its kinetic energy:

C = \frac{r\cdot v^{2}}{\mu }

From the NASA data table, r=6711.964km, and v=10.8343km/s at TLI. But what about μ? My JPL link from earlier in this post gives a value of 398600.435507km3/s2. (In the Apollo era NASA used an estimate that was a little different, but only after about the sixth significant figure.)

So for Apollo 11’s TLI:

C = 1.9766

We can now calculate orbital parameters a, e and θ. The semimajor axis depends only on C:

a=\frac{r}{2-C}

a = 286545 km

The eccentricity and true anomaly also depend on the flight-path angle (f):

e=\sqrt{\left ( C-1 \right )^{2}\cdot cos^{2}\left ( f \right )+sin^{2}\left ( f \right )}

e = 0.976966

\theta =atan\left ( \frac{C\cdot cos\left ( f \right )\cdot sin\left ( f \right )}{C\cdot cos^{2}\left ( f \right )-1} \right )

θ = 14.909º

The other parameters merely require a bit of spherical trigonometry:

i=acos\left ( cos\left ( \delta  \right )\cdot sin\left ( h \right ) \right )

i = 31.383º

Then we need a couple of intermediate calculations before we can derive our final two orbital parameters:

\varphi =asin\left ( \frac{sin\left ( \delta  \right )}{sin\left ( i \right )} \right )

To get this value into the correct quadrant before proceeding, we need to check the heading angle h. This is always between 0 and 180º for eastward launches (that is, pretty much every spacecraft launch). If it’s less than 90º we let φ stand, but if it’s over 90º, we set φ=(180º-φ). Then:

\omega =\varphi-\theta

ω = 4.410º

And:

\eta =acos\left ( \frac{cos\left ( \varphi  \right )}{cos\left ( \delta \right )} \right )

The longitude of the ascending node (Ω) depends on η, but also on which hemisphere we’re in, so we need to check δ before the final calculation. If δ<0º (ie, the southern hemisphere) then:

\Omega =\alpha +\eta 

Otherwise:

\Omega =\alpha -\eta 

Ω = 358.383º

And those are all our orbital parameters. Finally, we need an orbital period, which is:

P=\sqrt{\frac{4\cdot \pi ^{2}\cdot a^{3}}{\mu }}

P = 17.6679 days

Bringing it all together, we have a complete description of the orbit followed by Apollo 11 on its departure from Earth:

t = 2440419.18255527 days
a = 286545 km
e = 0.976966
i = 31.383º
Ω = 358.383º
ω = 4.410º
θ = 14.909º
P = 17.6679 days

In my second post I described how to extract a state vector from the Apollo data for re-entry, and we can plug those data into the same sequence of equations used above, to derive the elements of Apollo’s return orbit.

I could stop here, and you’re very welcome to stop at this point if you’ve managed to get this far. What follows is a description of how I went about checking that my calculations above produced valid results, and then a short Appendix that deals with calculating some alternative orbital parameters that you may need to know: the mean anomaly (M) and the time of pericentre passage (T).


We can get a quick sanity check on the calculations by looking at the summary “Translunar Injection Conditions” table from the Apollo 11 postflight trajectory document:

Apollo 11 Translunar Injection conditions
Click to enlarge (Source: Apollo/Saturn V Postflight Trajectory AS-506)

The orbital inclination figure from my calculation matches NASA’s exactly; the eccentricity goes astray in the sixth decimal place, presumably because of the difference between the 1960s estimate of μ and the present-day value. But if we add 180º to the descending node in NASA’s table we get an ascending node of 301.847º, which is a long way from my calculated value. But the reason for this is straightforward—NASA didn’t measure the position of the node in the conventional astronomical coordinates I’ve used to produce these orbital elements. Instead, they used a coordinate system defined by the onboard inertial navigation system housed in the Instrument Unit of the Saturn V launch vehicle. This system was kept constantly updated from the ground until just before launch, and the moment at which the inertial navigation system “locked in” its own set of coordinates (called Guidance Reference Release) was marked by a phrase many of us can remember from the Apollo launch sequence: “Guidance is internal!” This happened at time marker “T minus 17”—that is, 17 seconds before launch, and that marked the zero point for longitude, as far as Apollo was concerned.

The Julian Day corresponding to Guidance Reference Release for Apollo 11 is 2440419.06369213 UTC. If we use that to calculate GMST (remembering to go through UT1), we find that the Greenwich meridian was pointing towards a right ascension of 137.140º at that time. The launch pad for Apollo 11 (Complex 39 Pad A) is at longitude -80.604133º, and so at that instant had a right ascension of 56.536º (this is called the launch pad’s Local Mean Sidereal Time). That’s the point from which NASA measured its descending node for Apollo 11, and we need to add that angle to NASA’s nodes to convert to standard right ascension. I’ve already calculated that the ascending node in NASA’s coordinates is at 301.847º, which implies a right ascension of 56.536º+301.847º=358.383º. Which (ta-da!) is exactly the figure I calculated earlier.

These orbital elements are going to be pretty accurate for the first couple of hours of the translunar trajectory, during which the Apollo astronauts performed the Transposition, Docking, and Extraction manoeuvre, linking up with the Lunar Module and removing it from its cradle in the S-IVB stage. Shortly after that, they fired their engine for three seconds to kick themselves clear of the spent stage, acquiring an additional velocity of six metres per second—a pretty slight modification to the very high velocity at which they were leaving Earth behind. A day later, they fired the engine for another three-second burn in a midcourse correction manoeuvre—finely tweaking their trajectory to achieve a very accurate arrival in lunar orbit. And the orbit is going to be increasingly inaccurate as Apollo approaches the moon, when the moon’s gravity will progressively shift the trajectory away from the original ellipse defined by the TLI state vector.

But we can make another check on the accuracy of these orbital parameters, by checking if they at least deliver Apollo 11 to the vicinity of the moon at the time we know it entered lunar orbit—called Lunar Orbit Insertion (LOI), this happened at 17:21:50 GMT on 19 July 1969. Here’s what happens when I plug the calculated elements into Celestia, and set the time to LOI:

Apollo 11 orbit defined by TLI, extrapolated to the moon at time of Lunar Orbit Insertion - no midcourse corrections, no lunar gravity
Click to enlarge

Without any midcourse corrections, or any effect from lunar gravity, my simulated Apollo 11 shoots straight past the leading edge of the moon and is a short distance beyond it at the time of LOI. In reality, the moon’s gravity bent the trajectory around the back of the moon, where Apollo 11 fired its engine to slow down and enter lunar orbit. But this close rendezvous in simulation is a pretty good test that I’ve got all my orbital elements correct.

With all that in place, next time I come back to this topic I’ll write a bit more about Apollo 11’s first few hours on its departure orbit, and last few hours on its return orbit.


Appendix: For completeness, I should mention that there are other commonly used measures that define a spacecraft’s position in orbit at a precise instant—that is, alternatives to the true anomaly and epoch I derived above. The first of these is the mean anomaly (M) at epoch t. This is an angle calculated on the assumption the spacecraft moves around its orbit at constant angular speed, rather than accelerating and decelerating as it moves closer to and farther from the Earth. We get to M through an intermediary calculation of something called the eccentric anomaly (E).

E=acos\left ( \frac{1}{e}-\frac{r}{a\cdot e} \right )
M=E-e\cdot sin\left ( E \right )

For the Apollo 11 TLI state vector, this gives us:

M = 0.0375º

This is specific to the epoch t in the same way as the true anomaly.

Another parameter that’s often used is the time of pericentre passage (T). This abandons the epoch to which the state vector applies, and asks what the epoch would have been at the time the spacecraft made its closest approach to the body around which it orbits. And it doesn’t matter whether that close approach ever actually happened—there’s a close approach implicit in the maths of every elliptical orbit.

To come up with this number, we first convert M to a value measured in revolutions (divide the value for M given in degrees above by 360º) and then multiply be the period (P) in days. This tells us how many days ago the spacecraft would have passed through its closest approach. Then we subtract that number from the epoch (t) of the state vector, and we have the epoch (T) of the pericentre passage. In this case:

TTT = 2440419.18071554 days

And in this situation we don’t need to quote any sort of anomaly, because the anomaly at time T is (by definition) zero.


* Be slightly wary of online GMST calculators—many of them calculate GMST for noon on the Julian Day given, rather than for a specific time on that date.
This way of defining the longitude of the nodes, based on the Local Mean Sidereal Time of the launch site at Guidance Reference Release, applies for Apollo missions 10 to 17. The descending nodes quoted for Apollo missions 8 and 9 are based on Greenwich Mean Sidereal Time at Guidance Release. The mission reports for earlier flights give no descending node data.

The Advent Of Atomic Time

Apollo 11 launch
Click to enlarge
NASA image S69-39961

This is, to a large extent, a companion piece to my post about leap seconds, in which I described how the irregular rotation of the Earth means that the time as measured by our atomic clocks would fall out of synchrony with the actual movement of the sun in the sky, were it not for the occasional addition of a leap second. In this post, I’m going to look back at how the various systems of time measurement we inherited from the nineteenth century were forced to adjust to the advent of extremely accurate atomic clocks in the 1950s.

But this is also, as you might have guessed from the head photograph, relevant to my continuing project to derive Project Apollo orbital data. NASA’s early space programme was conducted during a period when time-keeping standards were in a state of flux, as I’ll describe, and that has implications for how we accurately specify the timing of significant orbital events, like translunar injection and atmospheric re-entry.

But first, something about the various timescales we use.

In the nineteenth century, time was a pretty straightforward thing. A day was the length of time it took the Earth to rotate on its axis relative to the sun. Because that duration varies a little during the course of a year, clocks were set according to the average position of the sun—to a “mean time”. And the mean time measured at the Greenwich Observatory in the UK was Greenwich Mean Time (GMT), which was adopted as an international standard at the Meridian Conference in Washington, in 1884.

This type of day, measured relative to the position of the sun, is strictly called a solar day, and it’s the only kind of day relevant to most people. So GMT was the basis for civil time—the time displayed on public clocks.

Astronomers are also interested in another type of day, however—the time it takes the Earth to rotate once on its axis relative to the fixed stars. This is called a sidereal day, and it’s about four minutes shorter than the solar day. Like the solar day, the sidereal day is measured in hours, minutes and seconds, but each of these measures is just 99.7% as long as the ones we’re used to. There’s a Greenwich Mean Sidereal Time (GMST) that describes the Earth’s rotational position relative to the stars, and that’s what tells astronomers in which direction they need to point their telescopes.

But astronomers are also interested in the solar day, if only because they need to know when the sun rises and sets. And for slightly complicated reasons that I’ve fudged straight past in the preceding three paragraphs*, they took to referring to their own version of GMT as Universal Time (UT) during the 1920s.

It turned out there was a certain irony to the “universal” bit of that designation, because since the nineteenth century there had been a suspicion that the Earth’s rate of rotation was not constant, and that therefore any definition of time based on the Earth’s rotation would be similarly inconstant. And so it proved to be—by the 1920s, it was evident that the movement of the moon and planets ran to a more steady timescale than the rotation of the Earth. While astronomers still needed Universal Time and Mean Sidereal Time, they also needed a more precise timescale against which to measure the dynamics of the solar system.

And so was born the idea of Dynamical Time—time calibrated by the observation of solar system events, particularly the movement of the moon. This was formally adopted under the name Ephemeris Time (ET) in 1952.

It was immediately evident that a clock ticking out Universal Time would diverge steadily from one marking Ephemeris Time—UT seconds were longer than ET seconds, because the Earth was rotating progressively more slowly. From the way in which Ephemeris Time was formally defined, it turned out that UT and ET had been in perfect agreement at some moment in 1902 but that, by the 1950s, the slower ticking of a UT clock meant that Universal Time was about 30 seconds behind Ephemeris Time, and that difference has been increasing almost ever since.

So we embarked on the 1950s with a set of timescales based on the rotation of the Earth (GMT, GMST, UT), and one based on the movement of the moon and planets (ET). Universal Time soon separated into three flavours: UT1, which tracked the Earth’s rotation; UT0, specific to certain local observations and affected by polar motion; and UT2, which is UT1 with a small correction applied to reproduce the predictable seasonal changes in Earth’s rotation. UT0 is no longer used, and won’t bother us here. UT2 has likewise fallen into disuse, but at the time I’m discussing was considered the gold standard for civil time-keeping, so will feature prominently in what follows.

But the time-keeping game changed forever during the 1950s, with the invention of the caesium atomic clock, which soon proved to be a more precise time-keeper than anything that could be achieved by the most careful astronomical observations. Given the variability of the UT second, atomic clocks were calibrated to tick off the standard second defined by Ephemeris Time.

The Système International (SI) system of units was a little slow to catch up with the advantages of atomic time. Until 1960, it continued to define the second as being 1/86400 of a solar day. Then it shifted to the Ephemeris Time definition, which was based on the Earth’s orbital period. And then, only seven years later, it switched to a definition of the second based on the calibration of the caesium atomic clock, which it has stuck with ever since.

It wasn’t until 1970 that we’d see a standard definition of atomic time, International Atomic Time (TAI), but TAI was just the culmination of series of other atomic time-scales used during the ’50s and ’60s, and continues seamlessly from them. As a result, we find that TAI was effectively synchronized with Universal Time (specifically, UT2) back in 1958.

Trouble was, of course, that Universal Time (and GMT) immediately started to drift away from the time kept by the atomic clocks. So during the 1960s we saw a struggle to come up with a way of somehow applying the extremely regular output of atomic clocks to the slippery and evolving timescale of the rotating Earth. This hybrid of atomic time and UT was eventually named Coordinated Universal Time (UTC). It consisted of a set of instructions issued by the Bureau International de l’Heure (BIH), telling people how to modify the output of an atomic clock in order to produce a time-signal that closely matched UT2. From 1961 to 1972, this consisted of a frequency shift and the occasional step-change of a twentieth or a tenth of a second, but in 1972 the BIH shifted to one-second steps designed to keep UTC within 0.9 seconds of UT1—the “leap seconds” I’ve previously written about. The BIH was dissolved in 1987, handing over leap-second duties to the International Earth Rotation Service, but UTC continues as the civil time standard applied around the world.

So that’s how atomic time became the basis for civil time. It soon also took over the role of Ephemeris Time. In 1976 an atomic standard called Terrestrial Dynamic Time (TDT) was synchronized with ET, and later renamed to just plain Terrestrial Time (TT). This is the timescale we currently use when figuring out orbital motions in the vicinity of the Earth. There’s another one, Barycentric Dynamical Time (TDB), used for calculating high-precision orbits in the rest of the solar system—it exists because of General Relativity, and is always within a couple of milliseconds of TT, so can often be neglected.

Because TAI and TT tick at the same rate, they bear a constant relationship to each other: TT=TAI + 32.184 seconds. Where does that offset come from? It’s because Ephemeris Time (uniform with TT) was synchronized with UT back in 1902, whereas TAI was synchronized with UT in 1958. And during that time, UT had drifted away from ET by 32.184 seconds.

So nowadays our timescales look very different from the way they were in 1950. We have the Earth’s rotation, defining UT1, monitored by Very Long Baseline Interferometry, and reported by the International Earth Rotation Service. UT1 is interconvertible with GMST, so if we need to calculate GMST, we go through UT1. Civil times everywhere are based on UTC, which is an atomic timescale with added leap seconds to keep it close to the observed values of UT1. And the fine detail of solar system dynamics are calculated using TT and its associated atomic times.

If I draw a diagram of my discussion so far, you can see that the 1950s, ’60s and ’70s were a period of intense flux for time measurement.

The advent of atomic timekeeping
Click to enlarge

This has relevance to my long, slow project of extracting orbital data from NASA’s original documentation, because NASA’s engineers would have been navigating to the moon using Ephemeris Time (which we can retrospectively call Terrestrial Time, because the one is a continuation of the other), but the mission documentation gives times according to GMT, or a time zone derived from it—either the time at the Florida launch site, or at Mission Control in Texas. So to correctly derive my Apollo orbits, I need to convert from GMT to TT—just at a time when the relationship between GMT and TT was at its most complicated.

To add to the complexity, the name “Greenwich Mean Time” was sometimes applied to UT1 when used for navigational purposes, so there’s a potential ambiguity to NASA’s usage of “GMT” in its Apollo documentation—did they mean UT1 or the evolving UTC standard? After wading through a lot of documents, I eventually turned up an answer to that question in the hefty Proceedings of the Apollo Unified S-Band Conference, which took place at the Goddard Space Center in July 1965. In the chapter entitled “Apollo Precision Frequency Source And Time Standard”, by R.L. Granata, we’re told that:

The method of obtaining time synchronization is to employ the WWV, HF signals.

WWV is the radio station used by the National Institute of Standards and Technology to broadcast time and frequency standards for the USA. These broadcasts were coordinated internationally by the BIH during the 1960s, and by 1965 had already been tied to the A3 atomic timescale, which was a direct precursor of TAI—so WWV was broadcasting the atomic-based civil time that would soon be known as UTC, and that is almost certainly what is meant by “GMT” in the Apollo documentation.

How do I get from UTC to TT? I need to turn to the Earth Orientation Center at the Paris Observatory, who maintain a dataset called the Earth Orientation Parameters, series C04. This provides daily values for UT1-TAI and UT1-UTC, stretching back to 1962. Bearing in mind the fixed relationship between TT and TAI, this is all I need to create a graph showing how UT1 and UTC were drifting away from TT during the course of the manned Apollo missions:

Time scales during the manned Apollo missions, 1968-72
Click to enlarge

You can see how, by dint of frequent small step changes and adjustments in clock rate, the BIH kept UTC extremely close to UT1 and UT2 right up to the start of 1972, which was when the leap second was introduced—a small step change at the start of that year brought UTC to exactly 10 seconds away from TAI (42.184 seconds from TT), and the first leap second then occurred at the start of July.

Pulling up the data for 16 July 1969, the date on which Apollo 11 launched, the Paris Observatory tells me that:

UT1-TAI = -7.5505119 s
UT1-UTC = 0.0115221 s

And we know that

TT-TAI = 32.184 s

So:

(TT-TAI)(UT1-TAI) + (UT1-UTC) = TT-UTC = 39.746 s

This value is sometimes symbolized ΔT, and it’s our route to converting NASA’s quoted GMT times to Ephemeris Time, or TT. And for strict accuracy, we also need to take note of the value for UT1-UTC, sometimes called ΔUT1. This is the conversion from UTC to UT1, and thence to GMST, which will be needed when I’m converting NASA’s state vectors to orbital elements in my next post on this topic.

So when NASA tells us Apollo 11 launched at 13:32:00 GMT, we can say that’s equivalent to 13:32:39.746 TT. It’s a significant difference—if we neglect it, it’s equivalent to a 16 kilometre displacement of the launch pad, and a 40 kilometre displacement of the moon. By comparison the conversion to 13:32:00.012 UT1 is trivial, and could easily be ignored, given the uncertainties in other data. More about that next time.


* Not like me, I know. If you want to know more about how the name “Greenwich Mean Time” came to be confusingly applied to several slightly different timescales, forcing the astronomers to give their own version a unique name, take a look at Dennis McCarthy’s “Evolution of Timescales from Astronomy to Physical Metrology”.
The progressive divergence of Universal Time stalled and reversed itself during 2020—solar days are getting shorter again, for reasons that are unclear.

Pole Stars Of Other Planets?

When I wrote about Philip Latham’s juvenile science-fiction novel Missing Men Of Saturn (1953) recently, I pointed out that Latham had made an astronomically well-informed guess about a possible pole star for Saturn’s moon Titan. Latham (a professional astronomer) knew the orientation of Saturn’s rotation axis, which would have allowed him to deduce the location on the celestial sphere around which stars would appear to rotate in Saturn’s sky, in the same way they appear to rotate around our pole star, Polaris, in Earth’s sky. And it was a reasonable guess that Titan’s rotation axis would be moderately well aligned with Saturn’s, leading Latham to have his protagonist use the star Gamma Cephei to orientate himself during his exploration of Titan.

So I thought this time I’d write about the location of the celestial poles of other planets—that is, the two apparently stationary points on the celestial sphere around which the sky of each planet appears to rotate, as a result of the planet’s rotation. Whether or not a star will turn out to be close enough to these points to function as a “pole star” is another matter. We’re in fact astronomically lucky (literally) to have a bright star situated so close to Earth’s north celestial pole; there’s no corresponding star in the south. And, because of the slow precession of Earth’s rotation axis, the alignment with the star Polaris is only a temporary one.* At present the celestial pole is moving slowly closer to Polaris, but Jean Meeus tells us, in his book Mathematical Astronomy Morsels (1998), that this motion will soon end, and pole and star will start to drift apart again—the closest approach will occur in February 2102, which is pretty close to being tomorrow, in astronomical terms.

Latham’s assessment that a major moon of a gas giant would tend to have its rotation axis aligned with its parent planet has proved to be correct. We now know that the major satellites of Saturn are so aligned (out as far as Titan), as are those of Jupiter, Uranus and Neptune. Even the two tiny satellites of Mars share their rotation axis with their parent. So in what follows, you can consider that the pole position of a planet, as marked on my small-scale sky maps, also indicates the pole positions of many of its satellites.

But before showing you my sky maps of the various celestial poles, I need to clarify what is meant by “north pole” and “south pole”.

If we look at the Earth from above its north pole, it appears to rotate anticlockwise, like this:

This gives us two, not entirely consistent, ways of defining “north” for other planets. The first would be to define the north pole as being that rotation pole from above which the planet appears to rotate anticlockwise. This is sometimes called the “right-hand rule”, because if we imagine wrapping our right hand around the planet’s equator, with our fingers pointing in the direction of its rotation, then a “thumbs up” sign from this position will point towards the planet’s north pole. This is a nice generalizable rule, and I’ll come back to it at the end of this post, but it’s not the one adopted by the International Astronomical Union.

In 1970, the IAU defined a sort of “solar system north”, using the Earth’s north pole as the criterion. The hemisphere of sky that lies on the north side of the plane of the solar system, as judged by the orientation of Earth’s north pole, defines the north poles of all the other solar system planets and their major satellites. The “right-hand rule” and “solar system north” produce consistent results when a planet rotates anticlockwise as viewed from solar system north. But they produce opposite results if the planet appears to rotate clockwise when viewed from solar system north. The IAU defines such planets as having retrograde rotation, because they turn backwards when compared to the Earth. The opposite of retrograde is prograde or direct, and most (but not all) of the major bodies in the solar system have prograde rotation.

You might be wondering how the IAU defines “the plane of the solar system”. Though all the planets orbit in roughly the same plane, they interact gravitationally with each other, so the precise angles between the various orbits varies. But the total angular momentum of the solar system stays the same, and that defines an axis and corresponding plane called (appropriately enough) the Invariable Plane of the Solar System (henceforth, the IPSS).

So there’s a “north pole” in the northern sky associated with the IPSS, and it, with its southern counterpart, are the only poles that have an absolutely constant position on an astronomical time scale. Nearby is another north pole associated with the plane of the Earth’s orbit around the Sun—the ecliptic pole. It’s near the IPSS pole because Earth’s orbit is tilted only slightly relative to the IPSS. It will move very slowly as the Earth’s orbit slowly shifts under the gravitational disturbance of the giant planets. And we know that the Earth’s axis of rotation is tilted at 23½º relative to the plane of its orbit, so the Earth’s north celestial pole lies 23½º away from the ecliptic pole, in the constellation of Ursa Minor, near the star Polaris. As I’ve already mentioned, this pole too is in motion—it describes a wide circle around the ecliptic pole every 26,000 years.

The north IPSS pole and ecliptic pole lie in the constellation of Draco, within the curve of the “neck” of the imaginary dragon that is sketched out by the constellation’s stars. A little cluster of other north poles is gathered around them. Venus and Jupiter follow orbits that are minimally tilted relative to the IPSS, and have rotation axes that are only slightly inclined to their orbits, so their poles sit close to the IPSS pole. Mercury also has a minimal tilt relative to its orbital plane, but its orbital plane is tilted by more than six degrees relative to the IPSS, moving its pole a corresponding distance from the IPSS pole. In the constellation diagram below, I’ve also marked the north pole of the Sun, which is tilted by around six degrees relative to the IPSS. So all the north poles mentioned so far lie in Draco, in or around the loop of the dragon’s neck, but none of them near any particularly bright stars.

On the maps that follow, planets with direct rotation are marked in green, and retrograde rotators in red.

North celestial poles of various solar system bodies
Click to enlarge

(You’ll notice that I’ve deviated from my habit of using Celestia to generate my astronomical and astronautical illustrations—all my plots this time are superimposed on star charts generated at In-The-Sky.org, which are clearer for my purposes today.)

Also plotted above are the poles of Saturn, Mars and Neptune, all significantly tilted relative to their orbital planes and the IPSS. Saturn’s north pole lies in a dark corner of Cepheus, actually a little closer to Polaris than to Gamma Cephei (marked with its name Errai on the chart). Mars has its pole inconveniently placed in the dark sky between Cepheus and Cygnus. Neptune’s pole has perhaps the most navigationally convenient location, on the left wing of Cygnus the Swan, almost on the line between second-magnitude Sadr and third-magnitude Delta Cygni.

In contrast to the other bodies plotted above, Earth’s Moon is marked with a dashed circle rather than as a single point. This is because, like Earth, the Moon’s rotation axis precesses around the ecliptic pole—but it does so in a mere 18.6 years. So on any given date the Moon’s north pole will be located somewhere around the circumference of a circle about three degrees across, centred on the ecliptic pole.

In the southern sky, the pattern of poles is the same, but inverted relative to their northern counterparts.

South celestial poles of various solar system bodies
Click to enlarge

The poles of Mars and Jupiter lie in dark parts of the constellations of Vela and Puppis, respectively. Mercury has a good, but relatively faint, southern pole star in third-magnitude Alpha Pictoris. Saturn’s pole star, Delta Octantis, is even fainter. This leaves the cluster of poles around the ecliptic pole, all of which lie in the faint constellation of Dorado. They all have a good pole “star” in the form of the Large Magellanic Cloud (not shown on my chart) which extends south from Delta Doradus and into the neighbouring constellation, Mensa.

One planet is missing from the charts above—Uranus. This is because Uranus notoriously orbits lying pretty much on its side, the inclination of its equatorial plane to its orbital plane being variously given as 98º or 82º and retrograde. This puts its poles in the vicinity of the plane of the ecliptic, rather than the ecliptic pole.

On my charts below, the brown line is the ecliptic, as a stand-in for the IPSS. Uranus’s south pole (retrograde rotation) lies in a dark area of sky, but is positioned between the two bright constellations of Orion and Taurus, making it relatively easy to find, at least approximately.

South celestial pole of Uranus
Click to enlarge

Its north pole lies in Ophiuchus, conveniently close to the second-magnitude star Sabik (Eta Ophiuchi).

North celestial pole of Uranus
Click to enlarge

Finally, I promised I’d come back to that “right-hand rule” definition. While the “solar system north” standard works well for planets and their large satellites, the International Astronomical Union realized that smaller bodies can have rotation axes that precess or migrate so quickly that they could easily shift back and forth across the IPSS within a few years. Under the “solar system north” rule, this would mean that the body would swap its north and south poles, and shift between direct and retrograde rotation, over the same period. So in 2009 they officially fell back on the good old “right-hand rule” for dwarf planets, minor planets, their satellites, and cometary nuclei. To avoid running two conflicting definitions of “north” in parallel, they designate the right-hand-rule pole of these bodies the positive pole, with the negative pole lying in the opposite direction.

Which brings us to Pluto, everyone’s favourite ex-planet. Until 2009 it was officially a retrograde rotator, with its north pole in Delphinus and its south pole in Hydra. Now, under the new definition, its positive pole is in Hydra, and its negative pole in Delphinus.

Location of positive celestial pole of Pluto
Click to enlarge
Location of negative celestial pole of Pluto
Click to enlarge

These poles will serve for Pluto’s huge moon, Charon, too. But sadly, both poles lie in rather dim and undistinguished parts of the sky.


* When the Phoenicians were setting out on their voyages of discovery, three millennia ago, they had no pole star to guide them. The north rotational pole of the Earth lay in a fairly empty bit of sky north of the bowl of the Little Dipper.

Which Way Does Space Station V Rotate?

Space Station V from 2001 A Space Odyssey

The “Phenomena” posts have been a little tied up with abstruse orbital mechanics and obscure revisions to lists of Scottish hills, of late, so I thought it might be time for a break from all that.

So this post is about something superficially trivial in the film 2001: A Space Odyssey, which has mildly annoyed me for the last fifty years.

People seem to particularly like compiling lists of mistakes in 2001: A Space Odyssey, presumably because it’s a classic film with pretensions to scientific accuracy, made by a famously exacting director advised by a famously knowledgeable science-fiction author. Who wouldn’t want to pick holes in that?

There are mistakes that are genuine scientific errors—I’ve written before about the dire physiological consequences that would have ensued if the Dave Bowman character had really tried to hold his breath when explosively decompressed. (Arthur C. Clarke later said that he would have advised against that, if he’d been present during the filming.) And there are mistakes that are simply technical malfunctions—like the glimpse of Bowman’s bare wrist we get when his spacesuit glove comes away from his spacesuit sleeve. And then there are “mistakes” which are self-evidently artistic decisions to step away from strict accuracy—the various planetary and solar alignments that herald great events, for instance, are clearly inconsistent with the astronomical positions of these bodies in previous scenes.

Finally, there are the continuity glitches, of which there are strikingly many in 2001, given what a notorious perfectionist Stanley Kubrick was.

So here’s the one that first caught my attention, when I watched the film from the front row of the top balcony of the Victoria cinema, back in 1970.

There are eight shots in which the giant wheel-shaped space station, Space Station V, appears in the film. I do love this thing, to the extent I’ve built a model of it. It rotates in order to generate “centrifugal gravity” at its rim, and it consists of two rings, one complete and one under construction. The docking port at the hub of the completed side is internally lit with white lights, while the inactive docking port on the other side is lit in warning red. So if we imagine approaching the station towards its active, white-lit hub (as the Orion III space shuttle in the film does), then we can state unambiguously whether it rotates clockwise or counterclockwise as seen from that vantage point.

When we first see it, twenty minutes in, to the accompaniment of Strauss’s Blue Danube waltz, it is rotating clockwise, by that definition:

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 1
Click to enlarge

I’ve marked the picture above with an arrow to indicate the rotation direction, and have noted the timing of the shot, in minutes and seconds, on my old “Deluxe Collector Set” DVD, from which I’ve made these muddy screengrabs.*

Then we get a series of shots involving the Orion III spaceplane on its way to the station, before seeing the station in the distance, with the approaching Orion III in the foreground and the moon beyond. Still clockwise.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 2
Click to enlarge

Next shot, another hero shot of the space station, with the pursuing Orion III appearing later. This is the point at which, back in 1970, I said (rather loudly, I’m told), “Hang on a minute!” The station is now rotating anticlockwise.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 3
Click to enlarge

Then a view from inside the cockpit of the Orion III. Still anticlockwise.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 4
Click to enlarge

There follows a brief view of the instrument panel, in which a wire-frame model of the docking port continues the anticlockwise rotation, and then in the next shot we’re inside the station hub looking outwards. The stars outside are moving in clockwise circles, implying the station is rotating anticlockwise as viewed from inside, and therefore (dammit) clockwise when viewed from our stipulated vantage point.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 5
Click to enlarge

Next shot, and we’re outside again, watching the spaceplane synchronize its rotation with the docking port. Anticlockwise again.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 6
Click to enlarge

Then a view from the Orion III cockpit, now synchronized with the docking port ahead. The stars sweep past in clockwise circles beyond the station, so the station (and spaceplane) are rotating anticlockwise.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 7
Click to enlarge

And this anticlockwise rotation is maintained in the final approach shot.

Rotation state of Space Station V, 2001: A Space Odyssey, Scene 8
Click to enlarge

So that’s just a minor curiosity—there are many worse errors in many other films, after all.

Except …

Once you start noticing this stuff, you keep on noticing it. 2001 is positively stuffed with left/right switches, as well as odd 90-degree anomalies. Another gross example occurs when we see the Earth, low on the Moon’s horizon. (This is another of those occasion on which artistry has overruled scientific accuracy—the Moon would be higher in the sky as seen from Clavius and Tycho, where the action takes place.)

As the Aries 1B shuttle approaches the Moon, we see the Earth illuminated from the right. But when we see a shot of several astronauts watching the shuttle’s approach, the illuminated portion faces left. Back to the shuttle, and it’s shifted right again. The same thing happens as we follow the moon-bus across the lunar terrain—the Earth starts off illuminated from the right, then switches to the left, then back to the right again.

The switching orientation of the Earth's illuminated limb in "2001: A Space Odyssey"
Click to enlarge

(In the Moon’s southern hemisphere, where the action takes place, the correct orientation at lunar sunrise, when the action takes place, is to have the Earth’s illuminated portion to the right, facing east.)

Later in the film, from shot to shot, we see the character Frank Poole not only reverse the direction in which he’s running around the Discovery centrifuge (twice), but the entire centrifuge (and Poole) becomes mirror-reflected:

The bone thrown aloft by the man-ape Moon-Watcher, at the beginning of the film, reverses its rotation direction between the two shots that follow its trajectory. And there are multiple geometrical inconsistencies during the sequence set on the Aries 1B moon shuttle, relating to the orientation of the control cabin.

It’s all very odd, and I don’t pretend to have an answer, but stuff like the constantly shifting orientation of the Earth seems too egregious to be anything other than deliberate—it would have been easier to have that not happen. And Kubrick, of course, has history with this sort of thing—the geometry of the Overlook Hotel in The Shining (1980) is notoriously protean.

Perhaps there’s a clue to what it all means, embedded in the very first shot in which we see a left-right reversal. The Dawn of Man chapter of the film famously used front-projected African scenes to provide back-drops for the outdoor sequences featuring the man-apes (which were actually filmed on a sound-stage). One particular reddened sky provided the backdrop for the set depicting the area around the man-apes’ cave refuge, and was used multiple times, both as a sunset and a sunrise. But in one shot (and only one shot), the sky image is reversed—and that’s in the shot in which the man-apes (and audience) first discover that a giant alien monolith has materialized outside the cave during the night.

Front-projected sky, Dawn of Man sequence, 2001 A Space Odyssey, showing mirror-reversal
Click to enlarge

* I’ve been trapped in a Tommy Lee Jones Cycle for decades, now. His character Kay in Men In Black (1997) kept having to buy the Beatles’ White Album over and over again, as physical media improved. I’ve now owned 2001: A Space Odyssey on VHS, DVD, Blu-ray and 4K disc. I suspect the rising graph of increasing visual and auditory fidelity from new technology has now crossed the descending graph of my failing eyesight and hearing.
I’m indebted to Juli Kearns for pointing out this key reflection in her shot-by-shot analysis of 2001: A Space Odyssey on her Idyllopus Press website.

Scottish Hill Lists: The Donald Revisions

Cover of 1953 edition of Munro's Tables

This is the second in my planned series of posts dealing with the revision history of the three “classic” tables of Scottish hills—the Munros, Donalds and Corbetts, which I introduced in an earlier post. I also introduced the idea of topographic prominence, and a way of charting these hill tables in two dimensions by plotting height against prominence. If any of this is strange to you, I refer you back to the original post via my link above, for a quick tutorial.

Last time, I dealt with the Corbetts (hills between 2500 and 3000 feet in height, with a prominence of at least 500 feet), and pointed out a number of ways in which new topographic data can lead to a hill either being deleted from, or added to, a set of tables.

This time, it’s the turn of the Donalds, lowland hills higher than 2000 feet. In contrast to Corbett’s tables, which have pretty simple and strictly numerical entry criteria, Donald’s tables feature a combination of rather more complicated topographic criteria with some value judgements, sorting the tabulated summits into two major categories—“Tops” and “Hills”:

“Tops”—All elevations with a drop of 100 feet on all sides and elevations of sufficient topographic merit with a drop of between 100 feet and 50 feet on all sides.
“Hills”—Groupings of “tops” into “hills” except where inapplicable on topographical grounds, is on the basis that “tops” are not more than 17 units from the main top of the “hill” to which they belong, where a unit is either 1/12 mile measured along the connecting ridge or one 50-feet contour between the lower “top” and its connecting col.

(Donald’s “drop on all sides” is the equivalent of modern “prominence”, the term I use in my charts below.) Donald’s rules seeks to reflect something about the shape of the landscape—allowing a single high summit, the “hill”, to dominate a fairly tight cluster of lower “tops”. The criteria given above mean that the summit of a “top” can’t be more than 1⅓ miles (2.15 km) from its parent “hill”.

In modern discussions of these tables, the “hills” have come to be referred to as “Donalds”, while the “tops” are called “Donald Tops”. The tables, as originally published, contained 86 Donalds and 47 Donald Tops. Donald also listed five English hills, close to the Scottish border, which fulfilled his criteria, but he did not assign them numerical entries in the tables. And, in an appendix, he added 15 summits:

[…] not meriting inclusion as tops, but all enclosed by an isolated 2,000-feet contour. These have been included in order that the table may be a complete record of every separate area of ground reaching the 2,000-feet level.

These locations have sometimes been referred to as “Minor Tops”, and that marginal category has actually been the main focus for such revisions as have been made, the remainder of Donald’s tabulation being surprisingly resistant to major change. Indeed, the Donalds remained entirely unrevised for 45 years, through multiple editions of Munro’s Tables.

Then, in the editions of 1981 and 1984, the availability of better mapping led to a considerable expansion to the list of Minor Tops, from 15 to 28—this despite the promotion of three Minor Tops (Keoch Rig, Conscleuch Head, and the south-west top of Windlestraw Law) to full Top status. These three were presumably selected on the basis of “sufficient topographic merit”, since they all have prominences between 50 and 100 feet.

The 1980s editions also ushered in a decade of confusion on the double-humped ridge of Black Law—creating a rather dubious Top on its north-east summit in 1981, to complement the existing Donald on the south-west summit; then switching the Donald and Top around in 1984, as the north-east summit proved to be higher than the south-west … only to have the Top deleted again in 1997, on the grounds (presumably) that its 36-foot prominence falls far short of Donald’s minimum criterion. So Black Law appears twice on my Donalds chart, with one summit marked as deleted and the other appearing as an addition. (A similar, later, migration of the Donald summit of Meikle Millyea is also marked. This was long anticipated, but not confirmed to the SMC’s satisfaction until 2015.)

Height-Prominence plot of original Donalds list, with revisions
Click to enlarge
Height-Prominence plot of original Donald Tops list, with revisions
Click to enlarge

The 1984 edition is also responsible for the only “promotion” of a Donald Top—Carlin’s Cairn.

Donald would have counted no less than seven fifty-foot contours on the ascent of Carlin’s Cairn, shown below on the one-inch mapping of 1926.

Ordnance Survey one-inch mapping of Corserine and Carlin's Cairn (1920s)
Click to enlarge

This whopping prominence, half again higher than any of the other Donald Tops in the original tables (see my plot of the Tops, above), means that Carlin’s Cairn meets Donald’s 17-unit criterion only because it’s less than a mile from nearby (and higher) Corserine. Donald was presumably swayed towards making it a Top rather than a Donald because it’s quite evidently part of the northern ridge of Corserine; but the 1984 Tables editor presumably felt that the comparatively large re-ascent from the col made the 17-unit rule “inapplicable on topographic grounds”, and so bumped Carlin’s Cairn to full Donald status.

Two significant revisions occurred in the 1997 Munro’s Tables. The first was the abandonment of the Minor Tops—they were either promoted to full Donald Top status, if merited, or deleted. Only one, Notman Law, survived the cull. (At the same time, Donald’s unnumbered list of five English summits was also dropped.)

The second revision was altogether more dramatic—the inclusion of a whole new and previously unsuspected group of Donalds and Tops. The discovery of these “Lost Donalds” on the south side of Glen Artney was first reported in The Angry Corrie, in 1994. Although Donald never clearly described what he meant by “the Scottish Lowlands” when he published his tables, it’s clear from the lists themselves that they document the 2000-footers of the Central Belt and Southern Uplands. The northern edge of this lowland area is commonly understood to be the Highland Boundary Fault (HBF). And this fault runs along Glen Artney, placing the 2000-foot hills on its southern side squarely in the Lowlands. You can check this for yourself on the Geological Survey of Great Britain (Scotland) Sheet 39W—Artney and the HBF lie in the top left corner and the new Donalds (Uamh Bheag and Beinn nan Eun) and associated Tops (Meall Clachach and Beinn Odhar) are visible in the Ordnance Survey mapping below the geological overlay. All are labelled on my charts.

Finally, there’s the vexatious (to me, at least) matter of Auchope Cairn and Cairn Hill West Top. The first of these two Tops was introduced by Donald, and discarded in 1997. The second appeared as a numbered Top in 1981 and is still with us. As my chart above shows, both fail to meet Donald’s 50-foot threshold prominence for inclusion, scoring 30 feet and a laughable 16 feet, respectively. Both are, also, much farther than the 17-unit threshold from the nearest Donald summit, at Windy Gyle; even The Cheviot, which featured as an unnumbered summit in Donald’s original tables, is not close enough to these two hills to play “Donald” to their “Top”. A map of the current Donalds and Tops makes their bizarre status in this regard clear:

Map of Donalds and Donald Tops
Click to enlarge

According to the 17-unit rule, the Donald Tops (open triangles) all lie within 1⅓ miles of their parent Donalds (filled triangles), forming dense clusters … except for Cairn Hill West Top, which sits in splendid isolation on the Scottish/English border. (Auchope Cairn is not marked, but lies only 700m north-west of CHWT, and would be superimposed upon it if plotted on my map.) Here’s the one-inch map of 1927, annotated with the position of Cairn Hill West Top (hereafter, CHWT):

Ordnance Survey one-inch mapping of The Cheviot border region (1920s)
Click to enlarge

So what’s going on? It appears that Donald was keen to include some indication of the highest point on the Scotland/England border, and Auchope Cairn was the closest named summit to that point. (There was a reluctance, in the early days of hill-tabulation, to include summits that were unnamed on Ordnance Survey maps.) The actual highest point was on the rounded shoulder of Cairn Hill, which has now been dubbed Cairn Hill West Top. This was marked by a 2422-foot spot-height on the Ordnance Survey six-inch map to which Donald would have referred; but that height had been inferred by sighting from a triangulation point about 800 feet to the south-west, with an altitude of 2419 feet.

So Donald provided Auchope Cairn with a rather gnomic footnote:

The highest point on the Union Boundary is (2,419) 2,422.

This footnote persisted until 1981, when the point was promoted to Top status under the newly minted name “Cairn Hill—West Top”. It was provided with a footnote that read:

The highest point on the Union Boundary. Not named on either O.S. or Bartholemew maps.

Auchope Cairn limped on in tandem to CHWT until it was finally deleted in 1997, presumably as surplus to requirements, leaving CHWT as an isolated anomaly—essentially a footnote with ideas above its station.


Note: My data source for this post is the Database of British and Irish Hills v17.2, combined with “The Donalds 1953-2021” dataset (version 3), both obtained from the DoBIH downloads page.

Scottish Hill Lists: The Corbett Revisions

Cover of 1953 edition of Munro's Tables

In a previous post, I wrote about the three “classic” Scottish hills lists—the Munros (1891), Donalds (1935) and Corbetts (1952), and how these were brought together, in a publication commonly referred to as Munro’s Tables, by the Scottish Mountaineering Club in 1953.

As a way of displaying the topographic data for these hills, I also introduced the idea of plotting each summit’s height above sea level against its prominence, a measure of its height above the surrounding terrain.

Height-prominence chart of Munros, Corbetts and Donalds
Click to enlarge

For more about the classic lists, the concept of prominence, and the design of the chart above, please refer back to my previous post.

It was inevitable that the classic tables would be overtaken by improved cartography, since they were based on early topographic surveys that have now been much improved upon. And although the idea of freezing these tables into historical documents has been discussed, particularly in the early days of Hugh Munro’s table of 3000-footers, there was also a countervailing idea that the compilers themselves would have embraced any changes imposed by improved cartography—Munro, for instance, continued to update his own tables throughout his life. So the SMC has “maintained” the tables, by sporadically publishing revised versions of Munro’s Tables and the associated guidebooks. (The pace of revision has slackened off in recent decades, as Ordnance Survey mapping has become more definitive, and the remaining “problem” hills have been subjected to careful survey with Differential GPS.)

What I’m going to do in this post (and two more) is to discuss the process of revision that has taken place. I’m going to do it in reverse chronological order, starting with the Corbetts and finishing with the Munros.

The Corbetts are a nice simple list to start with, since they’re based on well-defined criteria—a height between 2500 and 3000 feet, and a prominence of greater than 500 feet—so they occupy a very precise area of my height-prominence chart.

What I’ve done below is to plot Corbett’s original list of summits, but with the height and prominence we know they have today. Any original summits that are no longer part of the current tables are marked with a black cross; any summits in the current tables which were not listed by Corbett are marked with a red plus sign:

Height-Prominence plot of original Corbett list, with revisions
Click to enlarge

There are three obvious ways that Corbetts can end up being added to, or removed from, the tables. Firstly, a survey can show that a Corbett actually attains a height of more than 3000 feet, moving it into the “Munro” territory of the chart; or a hill previously considered to be a Munro can turn out to be lower than 3000 feet, potentially qualifying as a Corbett. So I’ve marked examples of hills that have crossed the 3000-foot divide since Corbett’s original compilation. Ruadh Stac Mor officially graduated to Munro status in 1974; Beinn Teallach in 1990. Beinn an Lochain moved the other way in 1974.

Secondly, we can have similar transitions at the 2500-foot limit of the Corbetts. Again, I’ve marked examples—Cook’s Cairn was “demoted” in 1990; Beinn na h-Uamha graduated to Corbett status as recently as 2016.

Thirdly, hills can make the transition in or out of Corbett status if a survey carries them across the 500-foot prominence line. This has been a relatively common way in which we’ve lost and gained Corbetts, primarily because prominence has been historically harder to pin down, since the Ordnance Survey understandably devoted more attention to finding the altitude of summits than defining the lowest point of cols. The transitions at this boundary are too many to label clearly, but you can easily see the cluster of crosses and pluses on either side of the 500-foot prominence line. Most of these transitions occurred in the 1981 and 1984 editions of the Tables, in the light of improved mapping.

But what about those deletions that have extremely low prominence? The deletion I’ve marked as “Sgurr nan Eugallt (East Top)” has a prominence of only 87 feet. Surely the Ordnance Survey could never have mapped that as exceeding 500 feet?

Here’s the mapping situation when Corbett was compiling his list—below is the relevant bit of the Ordnance Survey’s one-inch “Popular” edition, published around 1950:

One-inch "Popular" map of Sgurr nan Eugallt c.1950
Click to enlarge

You can see that the summit of Sgurr nan Eugallt, as labelled, is surrounded by a loop of 2900-foot contour—this is the summit that Corbett originally listed in his tables, with a height of 2933 feet.* But to the northwest there’s a broad rounded dome, also surrounded by a loop of 2900-foot contour, to which no-one seems to have paid any attention for fifty years. Corbett’s original summit appeared in every edition of Munro’s Tables up to the most recent, in 1997. But then in 2002 the second edition of the SMC’s guide-book The Corbetts & Other Scottish Hills suddenly pointed out:

Note that the true summit lies 600 metres or so NW along the undulating ridge.

According to more recent surveys, that broad rounded dome turns out to rise to 898 metres (2946 feet), whereas Corbett’s original summit comes in at only 895 metres (2936 feet). So the name Sgurr nan Eugallt has now been moved to a new home 600 metres northwest, while Corbett’s original summit is relegated to being merely “Sgurr nan Eugallt (East Top)”, with its prominence measured only from the nearby col. I’ve marked both summits in my chart. The pair Meall Coire nan Saobhaidh and Meall na h-Eilde have undergone a similar transition, with the former originally being considered the higher of a pair of two neighbouring lumps, but the honour moving to the latter in 1981.

So the Corbetts illustrate five potential ways in which a new topographic survey can change a hill’s status—too high, too low, insufficiently prominent, more prominent than previously thought, and turning out to be lower than a nearby summit to which the honour is transferred.

But my chart doesn’t capture the full complexity of the revision history of the Corbetts—some summits have made double transitions. For example, Corbett originally listed Sgurr nan Ceannaichean with a height of 2986 feet. Then in 1981 it was bumped to Munro status, with a listed height of 915 metres (3002 feet), only to be demoted again in 2009 when a more accurate survey revealed a height of 2997 feet.

Whereas Beinn Talaidh on Mull has made the opposite journey. In 1952 the Ordnance Survey showed it falling short of Corbett’s lower threshold by just four feet. In 1981 it popped up in the revised Corbetts list, with a note clarifying that:

Highest point lies 25 metres south west of the [triangulation] pillar and is 2502 ft.

But by the 1997 revision it had fallen off the Corbetts list again, with a height of just 761 metres (2497 feet).

So that’s the Corbetts— which were a nice, well-defined group to start with, illustrating most of the considerations that drive table revisions. Next time I’ll deal with the Donalds, which are complicated by being divided into two categories, Hills and Tops.


Note: My data source for this post is the Database of British and Irish Hills v17.2, combined with “The Corbetts 1953-2016” dataset (version 4), both obtained from the DoBIH downloads page.


* You’ll see that the one-inch map I’ve reproduced is inconsistently marked. The height “2933” appears to refer to a spot-height in the col, rather than to the summit marked Sgurr nan Eugallt, but this spot-height lies below the 2900-foot contour. Larger-scale maps (to which Corbett would have referred) clearly place the 2933-foot spot-height at NG 931044, on the summit originally marked as Sgurr nan Eugallt, with the col dropping to 2894 feet at NG 928046. Interestingly, the old six-inch map of 1902, which shows spot-heights but no contours, plots a spot-height of 2941 feet at NG 927048, on what we now understand to be the “real” summit of Sgurr nan Eugallt! So either Corbett missed this, or it was not present on the maps he consulted.

OS six-inch map of Sgurr nan Eugallt, 1902
Click to enlarge

Finding Apollo Trajectory Data

Apollo 11 Translunar trajectory relative to Van Allen Belts 3
Click to enlarge
Prepared using Celestia

A while ago I wrote a post entitled “How Apollo Got To The Moon”, which featured a few orbit graphics generated in Celestia, like the one above (which shows the orientation of Apollo 11’s departure orbit relative to the most intense region of the Van Allen Radiation Belt). I got a few enquiries about the data I’d used to plot the spacecraft orbits, and this is a long-delayed response to those enquiries.

As a preamble to this post, see my post “Keplerian Orbital Elements”, which describes in detail the various numbers used to define an orbit’s size, shape and orientation—the numbers I needed to plug into Celestia in order to draw the orbit in the illustration above.

This post is going to be about the NASA data sources I used, which don’t provide the orbital elements directly, but instead provide tables of state vectors, specifying the spacecraft’s position and velocity at a series of times. To plot the orbits the Apollo missions followed when they departed from, and returned to, the Earth, I need two specific state vectors. For departure, I want the state vector at Translunar Injection (TLI), which was defined to take place ten seconds after the Saturn S-IVB stage shut down its engines, having launched the Apollo stack on its way to the Moon. For the return orbit, I want the state vector for the Command Module at the atmospheric entry interface, defined as an altitude of 400,000 feet. In my next post on this topic, I’ll describe how to convert these state vectors into orbital elements.

A commonly used Apollo data source is Apollo By The Numbers (2000), a NASA publication compiled by Richard W. Orloff. Its data tables are widely available—they even have their own web pages on the NASA History website. The data I’m currently interested in appear under Translunar Injection and Entry, Splashdown, and Recovery. There’s also a useful page of Earth Orbit Data. But Apollo By The Numbers has the disadvantage of being a secondary source, containing a number of copying errors. So I went back to the primary sources—mission documents prepared immediately after each Apollo flight. These are available on-line as scans of the original typewritten pages, but the data are scattered across multiple sites, and links sometimes turn out to be broken. Even the NASA Technical Reports Server is missing some items, and is so full of oddly indexed material that it’s sometimes difficult to find even the material that is present. So I’ve spent some time recently compiling a collection of specifically Apollo-related documents at the Internet Archive, which I’ll link to as required below.

For the Apollo departure orbits, I took my TLI state vectors from the Postflight Trajectories prepared by Boeing. For atmospheric entry, my main source was the Mission Reports compiled at what’s now the Johnson Space Center.

The Postflight Trajectories include a pair of appendices charting reams of state vectors for the launch, Earth orbit and Translunar Injection phases on an almost second-by-second basis. The most useful tables for the TLI state vector are B-VII and C-VII (data in metric and imperial units, respectively). These have the look of being about as primary as you can get, given that they have the appearance of photo-reduced computer printouts, in contrast to the typewritten document to which they’ve been appended.

TLI state vector for Apollo 11
Click to enlarge (Source: Apollo/Saturn V Postflight Trajectory AS-506)

The relevant position data are the geocentric distance (GC DIST), longitude (LONG DEG E) and geocentric latitude (GC LAT DEG N)—a distance and two angles, which provide the spacecraft’s position relative to the centre of the rotating Earth. The relevant velocity data are the heading (HEAD DEG), flight-path angle (FLT-PATH DEG) and space-fixed velocity (SF-VEL)—again, two angles and a speed completely specifying the spacecraft’s velocity in three dimensions. The heading is the compass course, in degrees east of true north, along which the spacecraft is travelling; the flight-path angle is the angle above or below the local horizontal (parallel to the surface of the Earth directly below) in which it’s travelling; and the space-fixed velocity is its speed relative to a non-rotating Earth. This last value takes the spacecraft’s velocity relative to the surface of the Earth and augments it by the local rotation velocity of the Earth. That combined velocity determines how far the spacecraft will travel in its long elliptical orbit towards the moon—all the Apollo missions (and space missions in general) launched towards the east, to take advantage of the Earth’s rotation to give the spacecraft an extra boost.

These six numbers, together with the time and a knowledge of the Earth’s mass, are all that’s needed to derive an elliptical orbit that will be valid for the first few hours of Apollo’s departure from Earth.

But first we need to translate the time, given in seconds in the first column of the table above, into a format that’s meaningful in terms of orbital mechanics. The time given is what’s called the Range Time, or Ground Elapsed Time (GET)—the time since launch. Section 1 of the Postflight Trajectory report for Apollo 11 tells us that the mission launched on 16 July 1969, at 08:32:00 Eastern Standard Time. That corresponds to 13:32:00 Greenwich Mean Time. To make that date and time useful for plotting orbits, we need to convert it into a single number, the Julian Day (JD). My link gives you the necessary formulae to do that, but there are plenty of on-line calculators, too. There’s a suitable simple one here, which takes input in the form of the date and Greenwich Mean Time.* It only accepts a whole number of seconds, but all the Apollo missions launched on a whole number of seconds—most, like Apollo 11, on a whole number of minutes. I’ve nevertheless set the output to give the number of Julian Days to eight decimal places, to accommodate the TLI Range Time, which is quoted to a thousandth of a second in the table above.

The calculator tells us that launch (Range Zero) occurred on Julian Day 2440419.06388889. To that we need to add the Range Time of 10213.030 seconds (first dividing it by 86400, the number of seconds in a day). That gives us a total of 2440419.18209525 which, in the jargon of orbital mechanics, is the epoch of TLI.

The Postflight Trajectory documents also contain a summary table of the conditions at Translunar Injection, which include a smattering of Keplerian orbital elements—the inclination, descending node and eccentricity—which will provide good cross-checks on my own calculations.

Unfortunately, there’s no similar data source for the state vectors at the time of atmospheric entry—the Postflight Trajectory tables end with the start of the Transposition, Docking and Extraction manoeuvre, on the way to the Moon. The only reports consistently providing relevant data are the Mission Reports§, which span the time from launch to splashdown, and these seem to be the source for the tabulated data in Apollo By The Numbers. Here’s the table of entry conditions from the Apollo 11 Mission Report:

Apollo 11 entry trajectory parameters
Click to enlarge
(Source: Apollo 11 Mission Report)

The “miles” in the table are in fact nautical miles, the equivalent of 400,000 feet. The table provides a time, longitude, velocity, flight-path angle and heading angle, but lacks the geocentric latitude and distance that I need. Instead it gives a geodetic latitude and an altitude.

The time is the Range Time, again, and I can convert it to Julian Days in the same way I did for TLI above. The mission elapsed time of 195h03m05.7s corresponds to 8.127149 days, giving me an entry epoch of 2440427.191038.

To calculate the missing geocentric latitude and distance requires some research into the finer points of Apollo coordinate systems, and a bit of geometry. To find the geocentric latitude of a point, we (figuratively) draw a line from the point to the centre of the Earth, and measure the angle between that line and the plane of the equator. To find the geodetic latitude, we drop a line at right angles to the local horizontal plane, and measure the angle that makes with the plane of the equator. On a perfectly spherical Earth, these two latitudes would be exactly the same, but because the Earth bulges at the equator, they’re slightly different. Here’s a diagram, with the flattening of the Earth greatly exaggerated:

Diagram of geocentric and geodetic latitude of a spacecraft
Click to enlarge

The angle labelled ψ is the geocentric latitude, which is what we need for orbital mechanics; the angle labelled ϕ is the geodetic latitude, which is the latitude generally quoted in atlases and other geographical reference sources. The geocentric distance is the line r, connecting the centre of the Earth to the spacecraft, and the altitude is h, the distance between the surface of the Earth and the spacecraft, measured at right angles to the local horizontal plane.

Of course, the Earth’s surface isn’t a smoothly curving ellipsoid as in the diagram—but for the purpose of calculating geodetic latitude it’s treated as such. Various standard ellipsoids have been used by cartographers over the years, and by consulting the Project Apollo Coordinate System Standards, we can find out that the standard ellipsoid used throughout the Apollo missions was the Fischer “Mercury” Ellipsoid (1960). This ellipsoid was defined as having an equatorial radius (symbolized by a) of 6378166 metres. Its polar radius (b) was defined according to the ratio (ab)/a, called the flattening (f), which Irene Fischer determined to be 1/298.3. And that’s all I need in order to calculate the geocentric latitude and distance, using the geodetic latitude and altitude.

First, I need to work out the eccentricity (e) of the Fischer ellipsoid:

e=\sqrt{f\cdot\left ( 2-f \right )}

Then I need the length of R, which runs from the Earth’s axis to its surface in my diagram above, and is called the prime vertical radius.

R=\frac{a}{1- \left ( e\cdot sin\left ( \phi  \right )\right )^{2}}

The distance p, between the Earth’s axis and the spacecraft, measured parallel to the equator, is then:

p=\left ( R+h \right )cos\left ( \phi  \right )

And z, the distance between the equatorial plane and the spacecraft, measured parallel to the Earth’s axis, is:

z=\left [ R \left ( 1-e^{2} \right )+h \right ]sin(\phi )

Then the geocentric latitude is:

\psi =atan\left ( \frac{z}{p} \right )

And the geocentric distance is:

r=\sqrt{p^{2}+z^{2}}

Plugging in the Apollo 11 data from the table above, I get a geocentric latitude of -3.17 degrees, and a geocentric distance of 6500.02 kilometres. So now I have my state vector for atmospheric entry in the same format as for Translunar Injection, and I’m ready to calculate the orbital elements. Which I’ll do next time I return to this topic.


* Actually Universal Time, the successor to Greenwich Mean Time for astronomical time-keeping; but GMT was still the standard during the Apollo missions.
This is ludicrously optimistic, of course, given the uncertainty of some other numbers that will feed into the eventual calculation—but I prefer to do my rounding at the end, rather than in the middle.
Similar summary data appear in the Mission Reports, as well as in the Saturn V Launch Vehicle Flight Evaluation reports from what’s now the Marshall Space Flight Center. The Mission Report data generally seem to be derived from the Postflight Trajectories, though there are some departures from that. The Flight Evaluation data are often very slightly different from the Postflight Trajectory data—I think because of differing emphases on radar tracking and telemetry in the two reports from different organizations with different responsibilities. But none of the other sources provide anything as immediately useful as the state vector tables in the appendices to the Postflight Trajectories. (In the main, Apollo By The Numbers seems to take its data for TLI and Earth Parking Orbit from that source, though not entirely consistently.)
§ A number of earlier missions have Entry Postflight Analysis reports, which provide a more precise estimate of the entry state vectors than the Mission Reports, but this document was abandoned for Apollo 13 and doesn’t seem to have been reinstated.
This eccentricity is exactly the same property of the ellipsoid as the eccentricity that specifies the shape of an elliptical orbit, described in my post “Keplerian Orbital Elements”.

Scottish Hill Lists: The Classics

Cover of 1953 edition of Munro's Tables

If you’ve spent any time at all reading The Oikofuge, you’ll have gathered that I’m quite interested in hills—climbing them, looking at other hills from their summits, understanding their names and their place in history, landscape and land-use. What you won’t have seen me mention very often is the plethora of classifications that have been imposed on the Scottish hills over the years, starting with Sir Hugh Munro’s table of 3000-footers published in 1891, and culminating in the ongoing GPS-assisted activities of the good people over at the Database of British and Irish Hills.

It’s not that I’m uninterested in these tabulations, or the various parameters they’re based on. I have, after all, actually prepared a (long-obsolete) set of mountain tables all of my own. And the maps that accompany my various walk reports show the summits colour-coded according to their classification—you can find the key to the colours used in the FAQ section of the blog. And a glance at my annual CCCP reports will reveal a definite tendency to clamber up any 3000-footer that happens to be nearby, that being something of a raison d’être for the Crow Craigies Climbing Party.

But I don’t structure my walking activities around trying to “complete” any particular hill list—indeed, it’s only in the last few years I’ve attempted to reconstruct a list of all the summits I’ve visited in fifty years of hill-walking. And that process has led me to think a bit more about hill-lists in general, and how they came to exist. So I thought I’d write something about that. For this post, I’m going to start with the classics—the three Scottish hill-lists that dominated the mental landscape of hillwalkers back in the 1970s when I was first venturing out on to the summits.


Munros

Like the others in this trio, Munro’s list first appeared in the Scottish Mountaineering Club Journal, and has been curated by the Scottish Mountaineering Club ever since. His “Tables Giving All The Scottish Mountains Exceeding 3,000 Feet In Height” [SMCJ 1(6): 276-314] lists 538 Scottish “tops” that exceed the height limit. Munro then innocently instigated a century-long argument by separately enumerating those peaks he felt could “fairly be reckoned distinct mountains”. These were the 283 summits that came to be designated “Munros” in his honour; the remaining 255 tops on his list would then become the “Munro Tops”.

Munro worked on revising his list in the light of new mapping, and the SMC planned to issue his Revised Tables as part of their new General Guide-Book. This eventually saw the light of day in 1921, issued in instalments after a delay necessitated by the First World War—and, unfortunately, after Munro’s death in 1919. This list, consisting of 276 Munros and 267 Munro Tops, is perhaps the closest we can get to the “historical” Munros—Munro’s list largely devoid of input from other hands.

Unfortunately, Munro left no guidance on how he had decided whether one of his “tops” counted as a “mountain”. He certainly seems to have considered that large jagged mountains, like Beinn Eighe or An Teallach, could consist of only one “mountain” summit (the highest point) together with several mere “tops”; while he tended to scatter the “mountain” designation rather more profligately on rolling plateau land, like the Monadlaiths. This apparently unequal distribution of “mountains” relative to the difficulty of ascent would be the source of many later arguments, and I’ll come back to that when I write about the occasionally vexed topic of table revisions, in a later post. (There will be charts.)


Donalds

Munro’s 3000-footers are restricted entirely to the Scottish Highlands, and that may have been the inspiration for Percy Donald’s publication in 1935 of “Tables Giving All Hills In The Scottish Lowlands 2,000 Feet And Above” [SMCJ 20(120): 415-438]. This was a list of the highest summits in the Lowlands and Southern Uplands, thereby complementing Munro’s Highland-centric list.

After the pattern of Munro, Donald provided a list of 133 “tops”, and further classified 86 of these tops as “hills”. Unlike Munro, he attempted to provide some formal reasoning for his selection. But, again after the pattern of Munro, a degree of personal choice was permitted to creep in. His list of “tops” comprised:

All elevations [over 2000 feet] with a drop of 100 feet on all sides and elevations of sufficient topographical merit with a drop of between 100 feet and 50 feet on all sides.

Donald here introduces the idea of “drop”—the vertical distance between a hill’s summit and the highest connecting col. This is nowadays frequently called “topographic prominence”, and has become a key concept in modern hill lists. It can be formally defined as the summit’s height above the lowest contour which encircles the summit without enclosing any higher summit. But Donald then muddies the waters by making a subjective judgement with regard to the “topographical merit” of those 2000-foot eminences with drops in the 50-to-100-foot range.

In order to decide which of his “tops” were also “hills”, Donald gathered his tops together into groups, and nominated the highest of each group to be the “hill”, and the remainder to be “subsidiary tops” of that hill. His method of defining a group of tops is generally referred to by the SMC and other commentators as a “complicated formula”, but regular readers of The Oikofuge will no doubt recognize it as being a really simple formula. Donald measured the horizontal distance between adjacent tops along their connecting ridge, and measured the drop of the lower top by counting 50-foot contours between its col and summit. Each twelfth of a mile horizontally, and each 50-foot interval vertically, constituted one “unit”, and a “hill” could lay claim only to such subsidiary tops as fell within 17 units of its summit. As far as I can tell, Donald didn’t offer a justification for this particular formula, but it’s evident that his 17 units translate to about half-an-hour’s walking for someone setting a slightly more leisurely pace than the one stipulated by Naismith’s Rule. His “hills” are now called “Donalds” in his honour, and the subsidiary tops are “Donald Tops”. (In an appendix to his main tables, Donald also listed fifteen summits enclosed by isolated loops of 2000-foot contour, which are sometimes referred to as Minor Tops, but are of largely historical interest for reasons I’ll mention when I write about the revision history of these tables.)


Corbetts

The final member of the classic table trio (or triptych, as the SMC would no doubt style it) arrived in 1952, with the publication of J. Rooke Corbett’s awkwardly entitled “List Of Scottish Mountains 2,500 Feet And Under 3,000 Feet In Height” [SMCJ 25(143): 45-52]. To a considerable extent, this was a continuation of Corbett’s work tabulating the hills of England and Wales that rise to more than 2500 feet—a list he referred to as the “Twenty-Fives”, and had published in the Rucksack Club Journal in 1911.

Sadly, Corbett had died in 1949, and his tables were passed on to the SMC by his sister. The SMC appears to have been initially somewhat bemused, to judge from the foreword written by John Dow, who describes Corbett’s list of 219 summits as “incomplete”, stating that:

[…] reference to the maps—e.g., 1-in. Ordnance Sheets 42, 43, 49, etc.—makes it clear that numerous heights of equal “merit” to those listed have not been shown.

However, it soon became clear to the SMC that Corbett had in fact completed his tables—the apparent omissions were because he had, like Donald, applied a “drop threshold” below which summits failed to qualify for inclusion. Unlike Donald, he had not then applied any further, subjective judgements. When Corbett’s tables were republished in 1953, Dow’s revised foreword stated:

There was no indication in Corbett’s papers as to the criterion he adopted in listing the heights [ie, summits] included, but it seems clear that his only test was a re-ascent of 500 feet on all sides to every point admitted, no account being taken of distance or difficulty. No detailed check has been made, but the 500 feet qualification has obviously been exhaustively applied and rigidly adhered to […]

It is left as an exercise for the interested reader to figure out how this criterion could be deemed to have been “rigidly adhered to” in the absence of a “detailed check”, but a topographic prominence of 500 feet has been a stipulated qualification for Corbett-hood ever since.


These three tables were brought to together in a single publication in 1953: Munro’s Tables And Other Tables Of Lesser Heights on the cover, but more grandly styled Munro’s Tables Of The 3000-Feet Mountains Of Scotland, And Other Tables Of Lesser Heights on the title page. Its cover features at the head of this post.* It went through numerous editions and revisions over the course of the next four decades, until the most recent edition, in 1997, changed the title to the less judgemental Munro’s Tables And Other Tables Of Lower Hills. No matter: almost everyone refers to the publications as just “Munro’s Tables”.

A plot of the Munros, Corbetts and Donalds (according to the current lists) reveals some interesting features of their distribution:

Geographical distribution of Munros, Corbetts and Donalds
Click to enlarge

The Munros (in red) are confined by the nature of Scottish topography to the region north of the Highland Boundary Fault. Most are on the mainland, but two of the Inner Hebrides (Skye and Mull) host Munros. And we can see how Donald’s decision to confine his own tables to the Lowlands and Southern Uplands creates a complementary distribution of Donalds (in orange). Corbetts (in yellow) are spread all across Scotland—fringing the Munros in the north, reaching into several more islands, and mingling with the Donalds in the south. And because Donald set no upper limit to the height of his hills, there is in fact an overlap between the Corbetts and the Donalds—seven Donalds reach above 2500 feet with sufficient prominence to also qualify as Corbetts.

It’s informative, too, to plot the same hills on two axes according to their height and prominence:

Height-prominence chart of Munros, Corbetts and Donalds
Click to enlarge

In the absence of large areas of ground below sea level in Scotland, no hill can have a prominence greater than its summit’s height above sea level; and the only summits with prominence equal to their height are the highest points of islands. So I’ve plotted the Island Line on my chart, and labelled the three Munros and four Corbetts that lie on it.

The Corbetts cluster neatly, bounded by the 3000-foot contour above, the 2500-foot contour below, the Island Line to the right, and 500-foot prominence to the left. The Donalds sprawl a bit more—bounded by the 2000-foot contour below, but spilling into Corbett territory above, with seven orange triangles superimposed on the corresponding yellow Corbett plots. The Donalds all lie to the right of Donald’s 100-foot prominence cut-off; the Donald Tops (all bar one) lie to the right of his 50-foot lower limit. The reason for that anomalous Donald Top of negligible prominence will be explained (or at least, elucidated) when I write about later revisions to the tables.

Finally, the modern Munros list appears fairly well-behaved, too, with all the Munro Tops having prominences less than 500 feet, while the Munros themselves have prominences greater than 100 feet. This was not always so—it’s a product of later table revisions. And there’s another anomaly on the chart, in the form of a single Munro with negligible prominence. That something else I’ll explain in a later post on this topic.


Note: My data source for this post is the Database of British and Irish Hills v17.2, obtained from the DoBIH downloads page.

Resources: The original tables are slightly awkward to get at, being buried in large pdf scans of various volumes of the SMC Journal. And, once got at, the tables of Munro and Donald turn out to be difficult to read, the former having been printed in landscape orientation, the latter as double-page spreads. I’ve therefore prepared a little compendium of the relevant publications for these three sets of tables, rotating Munro’s landscape pages and merging Donald’s double pages for ease of consultation. The result is available on the Internet Archive here, to browse or download.


* It’s easy to be misled by the colophon that appears at the start of every edition of “Munro’s Tables” claiming a first edition in 1891, and two subsequent editions in 1921 and 1933. These dates refer to the first three publications of Hugh Munro’s tables, initially in the SMCJ and then in the two editions of the SMC’s General Guide-Book. The single publication commonly referred to as “Munro’s Tables”, containing the tables prepared by Munro, Corbett and Donald, didn’t (indeed couldn’t) come into existence until 1953, and I’m not sure why the SMC tries to push its publication history back into a time before the works of Donald and Corbett even existed.

This can lead to curious behaviour from walkers intent on “bagging” both Corbetts and Donalds. I once met a man on White Coomb who told me, in solemn tones, that he had been “forced” to climb the hill twice, because it was “once for the Donald and once for the Corbett”.

R.A.J. Matthews: Tumbling Toast, Murphy’s Law And The Fundamental Constants

Dropped toast

In what follows we model the tumbling toast problem as an example of a rigid, rough, homogeneous rectangular lamina, mass m, side 2a, falling from a rigid platform set a height h above the ground. We consider the dynamics of the toast from an initial state where its centre of gravity overhangs the table by a distance δ0

Robert A.J. Matthews published this seminal bit of applied physics in 1995. The journal reference is European Journal of Physics 16(4): 172-6, and you can access the full paper at ResearchGate, here. For his efforts, he was awarded an Ig Nobel Prize in 1996.

Matthews was the first (but by no means the last) to use mathematical physics to explore the popular claim that “dropped toast always lands butter-side down”. The usual “explanation” invoked for this perceived rule is Murphy’s Law—“If anything can go wrong, it will”—but Matthews sought to show that there were sound physical principles underlying the phenomenon.

He starts by dismissing the common physical explanations offered to account for this, principally airy claims relating to off-centre mass or aerodynamics effects created by the butter. He also dismisses those experiments that have claimed to disprove the rule—it’s unsurprising that buttered toast hurled randomly into the air* shows no particular preference for the side on which it alights, but this hardly reproduces the normal process by which toast falls.

Matthews starts with a static rectangle of toast, as described in the quotation at the head of this post. When its centre of mass moves beyond the edge of the table, it begins to tip over under the force of gravity. With any angle of tipping beyond zero (horizontal), gravity also produces a force that tries to slide the toast farther over the edge of the table. This is initially opposed by friction with the table edge, but eventually translates into a sliding motion. Gravity continues to accelerate the rate of rotation until the combination of sliding and rotation lifts the trailing part of the toast away from the table edge. Thereafter, the toast falls freely, and now rotates at a constant rate (neglecting air friction) until it hits the ground. If the toast rotates more than 90º but less than 270º on its way to the ground, it will strike butter-side down. Matthews appears to ignore the <90º regime during his initial analysis, presumably because toast falling from table height is observed to always rotate farther than that before hitting the floor.

In the quotation at the head of this post, Matthews sets the half-length of the toast to a, and the length by which the centre of gravity overhangs the edge of the table to δ. From these he defines an “overhang parameter”, η, equal to δ/a. The critical overhang parameter at which the tipping toast loses contact with the table edge is η0, and the tipping angle at which this occurs is φ. With g representing the acceleration due to gravity, he derives an equation for the constant angular velocity of the free-falling toast, ω0:

\omega _{0}^{2}=\left ( \frac{6g}{a} \right )\left ( \frac{\eta _{0}}{1+3\eta _{0}^{2}} \right )sin\phi 

The time, τ, it takes the toast to fall to the floor under gravity can be estimated using an approximation of the total distance it falls:

\tau =\sqrt{\frac{2(h-2a)}{g}}

And if the toast is to successfully rotate through “butter-side down” and into “butter-side up” during this time then:

\omega_{0}\tau > 270^{\circ }-\phi 

So that’s the story. Toast tips, slides, rotates free of the table edge, and then falls with a constant rate of rotation until it hits the floor after some elapsed time determined by the height from which it falls. If it rotates fast enough, or falls from high enough, it will manage to land butter-side up. But there will be a critical range of rotation rates and heights which will carry the toast into a butter-side-down impact.

The overhang parameter η0 is critical—if the toast has high enough friction with the table edge it will maintain contact with the edge for longer, allowing its rotational velocity to build up more before it falls free, maximizing the chance of a butter-side-up impact. Matthews derives a rather splendid formula for the minimum value of η0 which will generate sufficient rotational velocity for a butter-side-up landing.

\eta _{0}> \frac{2(h/a-2)\left ( 1-\sqrt{1-\frac{\pi ^4}{12(h/a-2)}} \right )}{\pi ^{2}}

(I’ve somewhat rearranged the equations in his paper, here, but the above is equivalent to those he provides.) For a table height h = 75cm and half-length of toast slice a = 5cm, it turns out that η0 has to be greater than 0.06.

Experiments involving bread, toast and kitchen Contiboard ensue, and Matthews finds that toast has a characteristic η0 of just 0.015, with untoasted bread only a little higher at 0.02. In his words:

This implies that laminae with either composition do not have sufficient angular rotation to land butter-side up following free-fall from a table-top. In other words, the material properties of slices of toast and bread and their size relative to the height of the typical table are such that, in the absence of any rebound phenomena, they lead to a distinct bias towards a butter-side down landing.

In fact, we can work out the minimum table height above which falling toast will have time to rotate far enough to land butter-side-up:

\frac{h}{a}=2+\frac{\pi ^{2}\left ( 1+3\eta _{0}^{2} \right )}{12\eta _{0}}

Plugging in the previously derived numbers yields an inconvenient minimum height of three metres.

Matthews then explores the effect of the horizontal velocity with which the toast departs the table edge—if fired over the edge with sufficient speed, the toast would have little time to start tipping over, would gain correspondingly little rotational velocity, and might stay relatively horizontal all the way to the floor. (That is, it would stay in the <90º rotation regime.) He concludes that the normal range of speeds with which toast is nudged off tables or tipped off plates is insufficiently high to prevent the butter-side-down landing.

Finally, there’s a section dealing with the fundamental constants of nature. In it, he builds on a paper by William H. Press, “Man’s size in terms of the fundamental constants” (American Journal of Physics, 48(8): 597-8), which you can find as a pdf here. Distilling down a more detailed argument, Matthews concludes that the upper height limit, LH, for humans is constrained by the ratio of the strengths of the electromagnetic force (which holds our bodies together) and the gravitational force (which breaks us if we fall from too great a height). If we got much taller than LH, we’d frequently sustain disabling or life-threatening injuries from simple trips and falls. After pushing around some equations, he concludes that:

L_{H}<\sim 50\times \left ( {\alpha /\alpha _{G}} \right )^{1/4}\alpha _{0}

Where α is the electromagnetic fine structure constant, αG the gravitational coupling constant for protons, and α0 the Bohr radius. These arguments are at best order-of-magnitude estimates, but Matthews plugs in the numbers and finds a surprisingly reasonable maximum figure of three metres for LH.

Matthews concludes that the frictional properties of toast set a limit on its rotation rate when falling from an edge, while the basic constants of the Universe set a limit on how tall humans are, which in turn sets a limit (about half LH) on how high useful tables are.

Our principal conclusion is a surprising one, given the apparently quotidian nature of the original phenomenon: all human-like organisms are destined to experience the ‘tumbling toast’ manifestation of Murphy’s Law because of the values of the fundamental constants of the universe. As such, we have probably confirmed the suspicions of many regarding the innate cussedness of the universe.

What to do? Reducing the size of the toast to match the scale of our tables is one solution, but the required size of ~2.5cm squares is (as Matthews remarks) “unsatisfactory”. He proposes instead the counterintuitive solution of speeding the toast on its way, to limit its opportunity to build up rotational velocity—flick it briskly over the edge, or snatch the supporting plate away, backwards and downwards.

So now you know.


* The BBC’s QED strand conducted just such an experiment in 1991.
Matthews’ work provoked a flurry of additional publications investigating the problem of tumbling toast. Analysis of video suggested that the free-falling toast rotates faster than Matthews predicted, probably because he had neglected the kinetic friction that occurs during the sliding phase. For more on the topic, take a look at the following:
Bacon ME, Heald G, James M. “A closer look at tumbling toast” American Journal of Physics (2001) 69(1): 38-43
Borghi R. “On the tumbling toast problem” European Journal of Physics (2012) 33: 1407-20