Accuracy Can Be Fun!

Last week, designer Thomas Romer of Chop Shop Studio launched a Kickstarter to support the creation of a new product: Planetary Blocks. I’ve been working closely with Thomas on these blocks for more than a year, advising on the content and researching the facts presented on some of the faces. I remember, as a kid, loving scientific diagrams like the periodic table of elements, or complicated-looking molecules, or evolutionary trees, or diagrams of plate boundaries. I didn’t understand most of what the numbers or symbols meant, but I remember staring at them, looking for patterns, trying to puzzle out their meaning. I’m hoping these blocks will inspire the same kind of curiosity in kids today.

We selected 20 worlds. The 8 planets and the Sun were an obvious first 9. Pluto also obviously had to be included, but it was very important to me that Pluto not be the only Kuiper belt object represented. That was tough, however, because the more of my favorite Kuiper belt objects we included, the fewer of my favorite moons and asteroids we could include. As for moons, we obviously needed our Moon, the 4 Galileans, Titan, and Triton. With 20 blocks, that left us with 3 more worlds to choose, at least 1 of which had to be another Kuiper belt object. We settled on Ceres as the only representative of the asteroid belt, Eris as Pluto’s companion, and added in Enceladus. It took Thomas some effort to sell me on Enceladus because there are so many other larger moons, and we didn’t have any of the Uranian moons, and that felt wrong. But I eventually agreed with Thomas that Enceladus is such a compelling exploration target that it’s important kids become familiar with it. Truth be told, I love all of Uranus’ moons, but if I’d had one or two more blocks I’d wouldn’t have added Titania or Miranda; I’d have added Sedna (so far away!) and/or Haumea (such a weird shape!). If I could do another 20 blocks, I would select, as a first cut: Pallas, Vesta, Ariel, Umbriel, Titania, Oberon, Tethys, Dione, Rhea, Iapetus, Charon, Makemake, Haumea, Sedna, 2007OR10, Orcus, Quaoar, 2002MS4, Salacia, and “worlds not yet discovered.”

Thomas had a pretty strong conceptual vision for these blocks, but I played a major role in developing the more fact-filled faces. Some of that was really hard work, but fascinating and fun nonetheless. It’s a pleasure to work with a designer who thinks that accuracy is as fun as I think it is! I thought you readers might enjoy some of the challenges I experienced as I helped with the development of this unique product.

Side 1: Names

Thomas wanted to put pronunciations of the world names on the blocks because some (Enceladus, Ganymede, etc.) are unfamiliar. The Planetary Society’s publications use Merriam-Webster for our dictionary, but their pronunciation guide employs phonetic characters that I thought would be too unfamiliar to many kids and parents. So I chose dictionary.com because their pronunciations are in plain Roman letters without requiring special font symbols. They’re less accurate, but more widely intelligible. I did make one editorial change: I went with the Merriam-Webster pronunciation for Io, styled in the Roman way dictionary.com would have written it (“AHY-oh”) instead of dictionary.com’s (“EE-oh”).

Side 2: Portraits

The art face depicts global views of each of the worlds, converted to a block-printed style by Thomas. We had a lot of back-and-forth on this one, whether the art side could depict more zoomed-in views of characteristic landforms, but I was concerned that if a kid didn’t have the corresponding photo in their mind’s eye, the block-print style representation of plumes on Enceladus or grooves on Ganymede would dissolve into a Rorschach test that wouldn’t mean anything. Thomas’ work on the global views is fabulous; I particularly liked the way he used a limited palette but still managed to make each world unique and recognizable. I drew on that inspiration for the composition side.

Side 3: Relative Sizes

How do you present the relative sizes of the Sun and other worlds of the solar system accurately to scale when every block is only 1.75 inches (4.4 centimeters) square? Thomas had a vision for this from the start, and it was great, essentially making a puzzle out of the whole solar system:

It was my job to get Thomas accurate numbers. For the most part, I was able to find diameters for the worlds on the JPL Solar System Dynamics website, largely from their Horizons service. JPL Horizons provides a handy table of facts for planets and Pluto, but you have to look up the characteristics of the moons individually through the Horizons Web interface. For reference, here are all the diameters I took from Horizons. (Horizons has them as radii, so I doubled those to get diameters since I figured diameters would be more meaningful to kids.)

• Mercury: 4,879 km
• Venus: 12,104 km
• Earth: 12,742 km
• The Moon: 3,475 km
• Mars: 6,779 km
• Jupiter: 139,822 km
• Io: 3,643 km
• Europa: 3,122 km
• Ganymede: 5,262 km
• Callisto: 4,821 km
• Saturn: 116,464 km
• Uranus: 50,724 km
• Neptune: 49,244 km
• Pluto: 2,377 km
• Enceladus: 504 km
• Titan: 5,149 km
• Triton: 2,707 km

For a couple of bodies, I needed to look up outside sources in order to get mean diameters:

• Ceres: 939 km (Park et al. 2016)
• Eris: 2,326 km (Sicardy et al. 2011)

Finding a good reference for the Sun’s diameter was surprisingly difficult, because it depends on your definition of where the edge of the Sun is. In the end I pulled my Encyclopedia of the Solar System, edited by Weissman, McFadden, and Johnson, off the shelf and looked it up in there: they give a diameter 1,391,940 km. Good enough for me.

Side 4: Orbital Positions

It was a huge challenge to communicate the relative positions and orbit sizes of all 20 worlds because of the variety of scales. I don’t know how many times Thomas revised the design of this face. I think his final version is remarkably information-dense, yet fairly clear. I feel like this is one of the sides I would sit there and puzzle and puzzle and puzzle over as a kid—I wouldn’t understand all of it but I’d line the blocks up and compare and think.

Unlike the comparative sizes face, Thomas and I didn’t try to communicate precise orbital distances; what’s important is to be able to compare distances from one world to the next. Because I didn’t need supreme accuracy, for most of the worlds, I could read the necessary information from the National Space Science Data Center (NSSDC). Here’s the NSSDC fact sheet for the planets. We made Mercury’s and Mars’ orbits slightly elliptical to match reality.

• Mercury: 0.387 AU
• Venus: 0.723 AU
• Earth: 1 AU
• Mars: 1.52 AU
• Jupiter: 5.20 AU
• Saturn: 9.58 AU
• Uranus: 19.20 AU
• Neptune: 30.05 AU

Moon distances are presented in lunar distances (LD), multiples of 384,000 km, calculated from the values given in the NSSDC fact sheet for large moons and Pluto.

• Moon: 1 LD
• Io: 422/384 = 1.10 LD
• Europa: 671/384 = 1.75 LD
• Ganymede: 1,070/384 = 2.79 LD
• Callisto: 1,883/384 = 4.90 LD
• Titan: 1,222/384 = 3.18 LD
• Triton: 355/384 = 0.92 LD

A few stray items come from JPL Horizons:

• Ceres: 2.77 AU
• Enceladus: 238/384 = 0.62 LD

We handled Pluto and Eris a little differently. Because their orbits are so elliptical, it seemed more important to represent that than to represent them at a specific distance from the Sun. So instead of giving Thomas a number, I took snapshots from the JPL Small-Body Database Browser that showed how the orbits relate to the nearly circular ones of the planets, and Thomas copied those. He moved the worlds along their orbits to a position straight out to the right of the Sun, so that their positions with respect to the Sun were parallel to his distance scales along the bottoms of the blocks.

Side 5: Missions

Originally, I proposed a puzzle for this side—a 4-by-5 array with an artwork of some kind on it, maybe a space exploration timeline or something. Unfortunately, Thomas had to abandon this idea for manufacturing reasons, and he proposed instead to present missions to each world. His art is, as always, great, but we had to make heartbreaking choices about which missions to list. I offered some suggestions but couldn’t pick favorites—they’re all my favorite!—so Thomas mostly wrote these sides. I love the art and will try not to think too hard about all the amazing missions that didn’t fit on this block face.

Side 6: Interiors

I felt very strongly that I wanted to do something to present what each world is made of, to invite comparison and contrast beyond the superficial appearances. My original concept was to do a graphical representation of the amount of each chemical element inside each world, but Thomas talked me down from that ledge. After much discussion, we settled on cutaway drawings showing and labeling internal layers. Here is where the rubber met the road in terms of both research and creative interpretation for me. How to generalize each world in such a way that kids can meaningfully compare one world to another?

The first step was for me to gather facts on what actually is inside each world. There’s no resource out there that presents cutaway views of more than a couple of worlds in one publication. I had to hit the library to find sources for each and every one. But there’s still a lot of guesswork and uncertainty involved in what we think these worlds are made of, because, of course, we can’t see inside the worlds directly. To understand how to draw these cutaway views, I had to understand the science.

There are several methods that planetary scientists use to figure out what is inside a world. The first is a very simple one: density. You can measure a planet’s size by looking at it, and measure its mass from a distance by timing the orbits of its satellites (or, if it doesn’t have satellites, by measuring its influence on other, more distant bodies). Mass and volume gives you density. Just knowing the density of a world tells you a lot about its composition, because there are 4 main substances that make up almost all the mass in the solar system: metal (mostly iron, with less nickel, and generally accompanied by sulfur because sulfur likes hanging out with metals); silicate rock (on average, the composition of iron- and magnesium-rich silicate minerals called olivine and pyroxene, accompanied by oxygen and the less-common metals in the periodic table); “ices” (mostly water with lesser amounts of ammonia and methane, and other rarer materials made of light elements like nitrogen and carbon); and gases (hydrogen and helium). If you know the density of a world, compare that to the densities of metal, rock, ice, or gas, and you have a first good approximation of what a world is made of. Earth is more dense than rock but less dense than metal, so it’s a mix of rock and metal. Pluto is more dense than ice but less dense than rock, so it’s a mix of ice and rock. Density is the main tool astronomers use to figure out what exoplanets are made of.

The next tool we have is gravity. Depending on how the mass is distributed within a world—if it’s the same kind of matter through and through, or if it’s layered with denser materials in the center—its gravitational effects on spacecraft will be different. We measure these effects by radio tracking of spacecraft flybys and orbits. Some worlds, like Ceres, Callisto, and Titan, seem to have pretty much the same density throughout, so the materials they’re made of are pretty uniformly mixed, without being strongly separated into layers. Other worlds, like our own Earth, are strongly differentiated, separated into density-driven layers: metal in the middle, then rock, then ice, then gas.

On most worlds, we don’t know exactly where these boundaries are. But on a few select worlds, we can use seismology to puzzle out not only where the boundaries are, but whether a layer is solid or liquid. On Earth, an enormous network of seismometers and high rates of geologic activity have allowed seismologists to identify global layering, the locations where rock gives way to metal, and to find out how much of the core is molten. We also have had seismometers working on the Moon, and recently put our first seismometer on Mars. Hopefully InSight will help us figure out where Mars’ core-mantle boundary is, and how much of its core is molten.

For a couple of other worlds, we have successfully tracked minute motions of features on the surface to deduce that there are internal molten layers that separate the outer shell of the planet from its core. We’ve seen that at Earth and Mercury (proof of their liquid outer cores) and Titan (proof of its liquid water ocean). On other worlds, notably the Galilean moons Europa, Ganymede, and Callisto, magnetic fields induced in saltwater oceans and picked up by orbiters have given away the presence of an ocean layer.

The last tool is math. Using differential equations or numerical simulations, a whole subfield of geophysics calculates the behavior of planetary interiors. Assuming an initial composition that matches what we know about the mass of a world, knowing how those materials conduct heat, knowing the rates at which planetary interiors generate heat through decay of radioactive elements, and making other reasonable assumptions, geophysicists can predict how a world is layered, even if we have no other insight into what’s beneath the surface.

Most of what we “know” about the interiors of worlds comes from physics. I put “know” in scare quotes because, in most cases, we haven’t tested these predictions yet. So I was a little fearful when I approached the challenge of telling kids what these worlds were made of through the medium of ultra-simplified blocks. But the point of the “core” faces of the blocks isn’t to teach kids specific details about each planet; it’s to facilitate comparing the planets, finding similarities and differences, sometimes in surprising places. (For instance: look at Europa and tell me whether you think it’s more similar to a terrestrial world like the Moon or Venus, or to an icy moon like Ganymede or Titan.)

And, of course, when pencil meets paper, artistic and pragmatic considerations enter, too. We couldn’t fit more than five layers into any block face. I chose to emphasize similarities over differences, so there are things lumped together that will probably make some planetary scientists wince. Let’s forge ahead anyway. Because I’m discussing layer thicknesses, I have to switch from talking about planetary diameters to planetary radii.

Mercury (radius: 2,440 km): As I was working on this article, a new paper on MESSENGER radio tracking by Antonio Genova and coauthors came out saying that Mercury has a solid inner core. That paper suggests a core-mantle boundary at about 1,970 km, with the inner core about half the radius of the outer core (so, 985 km thick each). The rest (470 km) is rock. There have been some other, earlier papers talking about a possible thin iron sulfide layer between the mantle and core. Just think—a kilometers-thick layer of iron pyrite! Fools’ gold in Mercury! The Genova paper considered this possibility but it’s not required to explain anything we know about Mercury, so reluctantly, I left it out of the drawing.

Venus (6,052 km): We know shockingly little about Venus’ interior—we don’t even know how concentrated its mass is towards its center—so all we have is physics. Venus is so big, however, that physics can get us pretty far. At Steve Hauck’s recommendation I consulted a book chapter by Frank Sohl and Gerry Schubert on “Interior Structure, Composition, and Mineralogy of the Terrestrial Planets” from the 2015 Treatise on Geophysics, Volume 10: Planets and Moons. Venus’ size means it retains a lot of heat so is certainly differentiated. Inside, it’s plausible that it looks roughly similar to Earth. Knowing its density and doing some math gives an estimated core radius of 3,050 km, and the remaining 3,002 km is therefore rock. We don’t know how much of the core is molten, so I just copied Earth’s molten proportion and called Venus’ solid inner core a radius of 1,050 km. I debated whether I ought to label some of Venus’ interior as partially molten, but decided I didn’t have a defensible source on that, so I left the whole mantle colored as solid rock, like Earth. On top of everything, I asked Thomas for a thin line to denote the atmosphere on top. That thin line is not proportional—it’s too thick, Venus’ atmosphere is only 80 km in height but it was important to me to note which worlds possess atmospheres, and which don’t.

Earth (6,371 km): This one, we know. From Sohl and Schubert, the core radius is 3,486 km, out of which the inner solid core is 1,217 km. The rest—2,885 km—is rock. Note that I’m not differentiating mantle and crust for any of the planets. Both are “rock,” for my purposes. On top of that, we add oceans and atmosphere. Thomas used some artistic license here—Earth’s oceans and atmosphere are even thinner than Venus’ atmosphere—but I can’t not include them. Thomas and I argued a little about whether Earth should be shown with active volcanoes like Io and Enceladus. Thomas’ position was that the continuous activity at Io and Enceladus is different in kind from Earth’s more sporadically erupting volcanoes, and also that Earth’s volcanoes aren’t connected with a molten interior reservoir, unlike Io and Enceladus. I’m still questioning this choice, but if I depicted volcanoes on Earth I’d also want to put them on Venus, and in both cases they’d have to sprout entirely above the atmosphere to be visible in the drawing, which would be wronger than not depicting them at all, so I acceded to Thomas.

The Moon (1,738 km): Researching this one really surprised me; I’d had the impression that the Moon was basically a solid, dead ball of rock. That’s wrong. According to Sohl and Schubert’s review, even though the Moon’s metal core is quite small (radius 450 km), it is mostly molten (solid inner core is only 100 km). Not only that, but even its deep rock is partially molten. I’m drawing in a partial melt layer of the first 450 km of the mantle, leaving the top 838 km solid rock.

Mars (3,390 km): From Sohl and Schubert, the core measures 1,794 km, leaving a 1,596-km-thick rock mantle. We don't know how much of the core is molten; figuring out that number is a goal of the InSight mission. I’ll just split it and say it's half molten until we get InSight results. As with Venus and Earth, we added a thin line for atmosphere.

Ceres (470 km): For this one I consulted a 2017 paper about Dawn results on Ceres’ interior from Roger Fu et al. It was difficult to generalize, and finally I just sent Fu an email to ask for help grossly simplifying the results. Mostly, Ceres is undifferentiated, made of mixed ice and rock. But there’s a teeny bit of layering near the surface, leftovers from a time long ago when there might even have been a surface ocean. That’s long gone, but Dawn observations suggest that there’s a region beneath Ceres’ surface where the temperature is warm enough and pressure low enough that some of those ices that are mixed with the rock are actually melted. So I’m calling the top 50 km icy rock, the next 50 km watery rock (colloquially known as mud), and the remainder down to the center, 370 km, icy rock. (According to the Fu et al. paper, the crust is actually weirder than that: it's 10% pore space, 25% water ice and clathrates, 30% clay silicate minerals, and 35% salty minerals. I'm ignoring the pore space and calling both kinds of minerals "rock" in order to arrive at "icy rock.")

Jupiter (69,911 km): For the giant planets I’m referring to the textbook An Introduction to the Solar System, edited by McBride and Gilmour, with some slight modifications based on more recent papers by Burkhard Militzer and coworkers. These are reeeeeeally schematic. Things inside giant planets are at intense pressure and very high temperature, and there are almost certainly not crisp boundaries between the layers as they’re drawn on the blocks, but I’m doing my best to describe what’s going on inside these giant worlds within the limitations of my medium. Jupiter probably has a core of rocky/icy materials about 5 times the mass of Earth—I’m calling that a 10,000 km radius. Next is a deep mixture of metallic hydrogen mixed with liquid helium that makes up the bulk of the planet—call that 50,000 kilometers deep. Next is a layer of liquid hydrogen and helium—I’ll give that 9000 kilometers. The final 911 kilometers is gas. I want to make very clear that these numbers are not precise, but they are at least representative, and that the main purpose for drawing them at all is for comparisons between Jupiter and Saturn. Because of space limitations, the blocks don’t mention the helium, just the hydrogen (which makes up 75% of the mixture).

Saturn (58,232 km): Saturn is made of the same stuff as Jupiter, but in different proportions because it’s smaller and the pressures in its interior aren’t as great. Evidence suggests it actually has a larger rock/ice core than Jupiter does, about 12 Earth masses, so I’m calling that a 15,000-kilometer radius. But its metallic hydrogen layer is only about half the rest of the planet; call it and the liquid hydrogen layer about 20,000 kilometers deep each. That leaves the remaining 3232 kilometers to be gas.

Uranus (25362 km) and Neptune (24622): For the ice giants, I developed models with advice from Amy Simon while working with her on her article for The Planetary Report about Uranus and Neptune. The two have roughly Earth-mass ice/rock cores (give those a radius of 5,000 km), over which there’s a layer of a weird material called superionic water. It has a lattice of oxygen atoms, so is solid in that sense, but hydrogen ions flow freely through it, like a liquid. I’m calling that layer a thickness of 14,000 km, which leaves the top 6,362 km of Uranus and 5,622 km of Neptune to be gas.

Io (1,822 km): For the Galilean moons of Jupiter, I referred to a 2004 book chapter by Schubert and coauthors, “Interior composition, structure and dynamics of the Galilean satellites” in Jupiter. The planet, satellites and magnetosphere. The thickness of Io’s core depends a lot on whether you assume it’s all iron or if most of Io’s sulfur is in there with the iron. This is a common issue for everyone trying to use physics to determine what’s inside a world. People seem to like the iron-plus-sulfur assumption, so I’m going with it and drawing Io’s core as 950 km thick. The core might be entirely molten, since Io lacks a dynamo-driven magnetic field. The mantle is probably partially molten, as much as 40%. There is a solid lithosphere atop everything, but it’s probably only 25 kilometers thick over a completely molten rock layer 200 kilometers thick. So our “block” model is 950 km molten core, 650 km partially molten rock mantle, 197 km molten rock asthenosphere, 25 km solid rock lithosphere.

Europa (1,565 km): Europa was another surprising one for me. It looks like an icy moon from the top, but slice it in half and it’s actually a terrestrial planet inside. This much we know from its relatively high density. But there’s no certainty about how its mass is distributed. We don’t know what minerals make up its rocky part, and those could incorporate more or less water, so we don’t know how thick any of its internal layers are. We also don’t know their states—melted or not (or partially so). Europa definitely has a liquid water ocean, but the ice shell thickness is also up for debate. I’m referring to a chapter from the 2009 University of Arizona Europa book by William Moore and Hauke Hussmann to draw its layers as 700-km-thick core, 740 km rocky mantle, and 125 km ice, of which the top 25 km is solid and the bottom 100 km is molten (that is to say, it’s a liquid water ocean). Based on a 2006 paper on Europa interior models by Fabio Cammarano et al, and making the same iron-plus-sulfur assumption I made for Io, I’m drawing the iron core as completely molten. I tried to figure out whether I should draw any amount of the rocky mantle as partially molten, but it’s too uncertain; I left it solid.

Ganymede (2,631 km): Ganymede is fun because it is so large (bigger than Mercury!) but it’s much less dense than Europa or Io. That means a lot of ice. Deep ice behaves in weird ways. Its ice goes so deep, to such high pressures, that it might have a really funky layered structure with more than one ocean layer separated by solid layers of ice of different crystal structures. Steve Vance calls this a “Dagwood sandwich” model of Ganymede’s interior. Based on Schubert et al. (2004), I’m drawing Ganymede with a 900-km core, an 831-km rocky mantle, and 900-km outer icy layer. Ganymede has an internally driven magnetic field, so its core has to be partially molten; I’m splitting it in half, so drawing a 450-km solid core and 450-km liquid outer core. Then, the Dagwood sandwich of liquid and solid water layers, which are fairly schematic—I didn’t try to dictate layer thicknesses there, only that they alternate.

Callisto (2,410 km): Callisto’s an interesting transitional world. It’s large, just barely smaller than Mercury, but smaller and colder than Ganymede and apparently has not differentiated much, because its mass is only slightly concentrated toward the center. And yet it has a liquid-water ocean perched within a water ice layer near the surface. Based on Schubert et al. (2004) I’m drawing it with an outer ice layer 125 km thick, then a 70-km ocean, then 45 km more ice, then ice-rock mixture for 1500 kilometers, over a rocky core with radius 670 km. In cross section it looks like Titan, minus the atmosphere. Or Ceres, plus an ocean and ice crust.

Enceladus (252 km): It may be really small, but according to a 2016 paper by Peter Thomas and coauthors, Enceladus has a global ocean. That ocean probably meets the rocky core at a depth of 50 km, globally. That gives a 202-km core. Above the ocean, there’s an icy crust that’s thinner (maybe 12 km thick) at the south pole, and thickens to maybe 25 km thick toward the equator.

Titan (2,575 km): Titan is Callisto-like, but the jury’s out on whether it has a rocky core or whether its interior is pretty much undifferentiated. There’s a 2010 paper by Julie Castillo-Rogez and Jonathan Lunine on possibilities for Titan’s interior indicating that either is a possibility. I’m drawing it undifferentiated. There is definitely an outer ice layer with an ocean layer perched in the middle of it. I’m drawing Titan with a 200-km outer ice layer, beneath which there is a 100-km ocean, then a 175-km ice layer, and the rest as rocky interior. Unlike Callisto, or actually any other solid-surfaced world, Titan has an atmosphere thick enough to register at the scale of these diagrams. It extends 200 km above the surface. Therefore, I told Thomas to render Titan with a diameter of 2,775 km and to draw in the atmospheric thickness proportionally. He added in the hydrocarbon lakes, which are nowhere near as thick as they are shown (their thickness is negligible at this scale), but as with Earth’s oceans I just had to include them because Titan has liquid lakes and that is cool and rare.

Triton (1,353 km): This far out in the solar system, there is not a lot of data. I’m relying on a fairly venerable (2006) paper by Hauke Hussmann et al. on “Subsurface oceans and deep interiors of medium-sized outer planet satellites and large trans-Neptunian objects.” It’s a physics paper that takes the known sizes and masses of these worlds and asks, first, how they would be layered if they were made of a rocky core and an icy crust, and second, whether there might be a water ocean at the base of the icy crust. They also explore a little variation in the composition (and therefore density) of rock and ice. These models are, at best, schematic, but it’s the best we can do for now. For Triton, Hussman et al. (2006)’s middle-of-the-road compositions produce a rocky core 1,017 km in diameter, above which there is an ocean of thickness 149 km, and then a 188-km ice crust. Hussmann’s paper includes predictions for tons of other solar system worlds not represented on this set of 20 blocks, so it provides an opportunity to expand the block set, if enough people support the Kickstarter...

Pluto (1,188 km): That 2006 Hussmann et al. paper was the best thing we had until New Horizons. New Horizons results from Bill McKinnon et al. (2017) show a rocky core 890 km thick and an ice shell 298 km thick. It’s possible that there’s an ocean at the bottom of the ice layer, so I’m drawing that 68 km thick, leaving the solid ice shell 230 km thick.

Eris (1,163 km): Hussmann et al. (2006) discuss Eris (then, it was still named 2003 UB313) but they didn’t know its correct diameter yet. Eris turned out to be highly reflective, so astronomers had overestimated its size and underestimated its density at the time the paper was written. So I’ll admit I just kind of guessed at this one. It’s almost as big as Pluto but denser, so I’m showing it with a rocky core 985 km thick, with an ice shell 125 km thick over a 53-km ocean.

Phew. Like I said, a lot of work, but it was a lot of fun to journey across the solar system in this way, diving beneath the surfaces of the planets, moons, and other worlds. Check out the Kickstarter if you’d like to support this project!

A previous version of this article misstated the radius of the Moon and consequently had incorrect values for the thicknesses of its internal layers.