Subsurface Oceans

An ocean with a roof, not a sky

Picture Earth's ocean, then freeze the top thirty kilometres of it solid and slide the whole thing out past Jupiter where the Sun is a bright star and nothing more. That is roughly the situation on Europa. Beneath an ice shell as thick as the deepest part of Earth's crust sits a globe-spanning ocean of salty liquid water — dark, pressurised, and never touched by sunlight. The water is liquid not because anything warms it from above but because the moon is being squeezed from outside and heated from below.

These hidden seas are not exotic edge cases. They are, by volume, where most of the liquid water in the Solar System actually is. The surface oceans of Earth are a rounding error next to the combined inventory of the icy moons. What makes a subsurface ocean possible is a deceptively simple recipe: a body big enough to start with some internal heat and to hold onto it, a thick insulating lid of ice on top, and — for most of them — a way of generating fresh heat continuously through tidal flexing. Where all three line up, an ocean can persist for billions of years in total darkness.

The defining feature, and the thing that separates a true ocean world from a merely icy one, is that the liquid layer is global and decoupled — the ice shell floats freely on top of it, mechanically disconnected from the rocky interior. That decoupling leaves fingerprints we can read from orbit: a wobble in the moon's rotation, a magnetic echo, a surface scarred by stress it could only feel if it were sliding on liquid.

The physics: why the water doesn't freeze

The central question is an energy balance. Heat flows out of the moon through the ice shell by conduction; if the interior cannot replace that lost heat, the ocean freezes solid from the bottom of the shell downward. The ocean survives only where the internal heating rate matches or exceeds the conductive loss.

The dominant source for the small, eccentric moons is tidal dissipation. A moon on an eccentric orbit feels a tidal force from its planet that varies over each orbit, both in strength (closer at periapsis, weaker at apoapsis) and in direction (the sub-planet point librates as the orbital angular velocity diverges from the constant spin rate). The resulting periodic deformation flexes the interior, and because real materials are not perfectly elastic, a fraction of that mechanical work is lost to friction as heat. The tidal heating rate scales as

Ė_tide = (21/2) · (k₂/Q) · (R⁵ n⁵ e²) / G

  k₂  = degree-2 tidal Love number (rigidity response)
  Q   = tidal quality factor (1/Q ∝ energy lost per cycle)
  R   = body radius
  n   = mean orbital motion (= 2π / orbital period)
  e   = orbital eccentricity
  G   = gravitational constant

The brutal sensitivities here are the keys to the whole subject. Heating goes as e² — double the eccentricity, quadruple the heat — and as n⁵, so a short orbital period (close-in moon) is enormously more effective. This is why Io, the innermost Galilean moon, is the most volcanically active body in the Solar System, and why Enceladus, deep in Saturn's gravity well, can sustain an ocean despite being only 504 km across.

But eccentricity should damp to zero as tides circularise the orbit. It does not, because of orbital resonance. Io, Europa and Ganymede are locked in the Laplace resonance, with orbital periods in a near-exact 1:2:4 ratio; their repeated mutual tugs continually pump eccentricity back up, sustaining the heating. Enceladus is held eccentric by a 2:1 resonance with Dione. Without these resonances the oceans would have frozen long ago.

For the larger moons (Ganymede, Callisto, Titan), radiogenic heating from the decay of long-lived isotopes (²³⁸U, ²³⁵U, ²³²Th, ⁴⁰K) in the rocky core dominates, supplemented by the latent heat released as the ocean slowly freezes and by antifreeze chemistry. Salts and especially ammonia depress the freezing point sharply — a water-ammonia eutectic freezes near 176 K (−97 °C) rather than 273 K — which is why Titan's deep ocean can stay liquid on a body whose surface sits at 94 K.

The key numbers

Real values anchor the picture. The Sun delivers about 1361 W/m² at Earth; at Europa it is roughly 50 W/m², and at Enceladus about 15 W/m² — far too little to matter for an ocean buried under ice. Internal heating dwarfs it.

World	Diameter	Ice shell	Ocean depth	Primary heat source	Surface temp
Europa	3,122 km	~15–25 km	~60–150 km	Tidal (Laplace 1:2:4)	~102 K
Enceladus	504 km	~5–25 km (1–5 at pole)	~10–40 km	Tidal (2:1 Dione)	~75 K
Ganymede	5,268 km	~100–150 km	~100 km (sandwiched)	Radiogenic + tidal	~110 K
Callisto	4,821 km	~100–150 km	~few × 10 km	Radiogenic	~134 K
Titan	5,150 km	~50–100 km	~few × 100 km	Radiogenic (NH₃ antifreeze)	~94 K
Pluto	2,377 km	~150–300 km	tens of km (likely)	Radiogenic + latent	~40 K

The pressures at the seafloor (or ice-floor) are formidable. On Europa the ice-ocean interface sits under tens of bars; at the base of a 100 km ocean the pressure approaches a kilobar — still in the regime where liquid water and ordinary ice (Ih) coexist. On the larger moons the ocean is thick enough that pressure at its base exceeds ~2 GPa, where water freezes not into ordinary ice but into dense high-pressure ice phases (ice V, VI). That high-pressure ice forms a floor, potentially sealing the liquid off from the rock — a key complication for habitability discussed below.

How we found them

You cannot see an ocean under 20 km of ice. Every detection is indirect, and the cleanest is electromagnetic induction. Jupiter's magnetic dipole is tilted ~10° from its rotation axis, so as the planet spins (every 9.9 hours) the field at a moon's location oscillates. A salty, electrically conducting ocean responds by Faraday's law: the changing external field drives eddy currents in the ocean, which generate a secondary, induced magnetic field opposing the change. The signature is a field that varies at Jupiter's synodic rotation period with a specific phase and amplitude set by the ocean's depth, thickness and conductivity.

NASA's Galileo orbiter measured exactly this at Europa during close flybys in 1998–2000: the induced field implied a global conducting layer (a salty ocean) within ~100–200 km of the surface. The same magnetometer found induction signatures at Ganymede and Callisto too. Ganymede is doubly special — it is the only moon with its own intrinsic, internally generated magnetic field, on top of which the ocean's induced signature rides.

The other gold-standard method is direct sampling. At Enceladus, the Cassini spacecraft discovered south-polar plumes in 2005 and then flew through them — most daringly in October 2015, skimming 49 km above the surface. Its instruments tasted the spray directly: water vapour and ice grains laced with sodium and potassium salts, silica nanoparticles (pointing to high-temperature water–rock chemistry at ~90 °C+), and complex organic molecules including species with masses above 200 atomic mass units, plus molecular hydrogen. The salts and silica are only explicable if the plume draws from liquid water in contact with rock — i.e., a real ocean with a hydrothermally active seafloor.

Supporting evidence comes from physical libration (Cassini measured Enceladus wobbling more than a solid body could, proving the shell floats free of the core), from gravity and shape (the moment of inertia constrains the layered interior), and from surface geology — Europa's chaotic terrain, double ridges and crater-poor youth all imply an active, mobile shell over liquid.

Worked example: does Europa's tidal heat keep its ocean liquid?

Let's check the energy balance with order-of-magnitude numbers. First the heat loss through the shell. Conductive flux through ice of thickness D with thermal conductivity k ≈ 2.5 W m⁻¹ K⁻¹ across a temperature drop ΔT from the ~273 K ocean to the ~102 K surface:

q = k · ΔT / D
  = 2.5 W/m/K × (273 − 102) K / 20,000 m
  ≈ 0.021 W/m²

Europa surface area  A = 4π R²  with R = 1,561 km
  A ≈ 3.06 × 10¹³ m²

Total conductive loss  P_loss = q · A
  ≈ 0.021 × 3.06 × 10¹³
  ≈ 6.5 × 10¹¹ W   (about 650 GW)

Now the tidal supply. Europa's eccentricity is e ≈ 0.009, orbital period 3.55 days (n ≈ 2.05 × 10⁻⁵ s⁻¹), radius R = 1.561 × 10⁶ m, and plausible interior values k₂ ≈ 0.25, Q ≈ 100. Plugging into the tidal heating formula:

Ė_tide = (21/2) · (k₂/Q) · R⁵ n⁵ e² / G

R⁵  ≈ (1.561e6)⁵  ≈ 9.3 × 10³⁰ m⁵
n⁵  ≈ (2.05e-5)⁵  ≈ 3.6 × 10⁻²⁴ s⁻⁵
e²  ≈ 8.1 × 10⁻⁵
k₂/Q ≈ 0.0025

Ė_tide ≈ 10.5 × 0.0025 × 9.3e30 × 3.6e-24 × 8.1e-5 / 6.67e-11
       ≈ 1 × 10¹² W   (order ~10¹¹–10¹² W)

The two numbers — a few hundred GW lost, of order 10¹¹–10¹² W supplied — are comparable. That rough balance is exactly why Europa sits in a steady state with a thin-to-moderate shell over a persistent ocean, rather than freezing solid or melting through. (The real calculation distributes dissipation between the ice shell and a partially molten rocky mantle, and the result is sensitive to the poorly known k₂/Q; this estimate just shows the budget closes.) For Io, push e and n up and the same formula yields ~10¹⁴ W — global volcanism instead of a buried ocean.

Discovery: predictions, then proof

The story begins with a prediction. In 1979, weeks before Voyager 1 reached Jupiter, Stan Peale, Patrick Cassen and Ray Reynolds published a paper in Science calculating that tidal heating should make Io intensely volcanically active. Voyager 1 promptly photographed erupting volcanoes — one of the great confirmed predictions in planetary science. The same logic implied Europa might harbour a liquid layer.

The decisive ocean evidence came from Galileo (orbiting Jupiter 1995–2003). Magnetometer data from Margaret Kivelson's team revealed Europa's induced field around 2000, establishing the salty ocean. Meanwhile Cassini (at Saturn 2004–2017) transformed Enceladus from a curiosity into the most accessible ocean world: the 2005 plume discovery, the 2014 gravity result implying a regional-to-global south-polar sea, the 2015 confirmation of a global ocean from libration, and the in-situ plume chemistry showing salts, silica and H₂.

The current era is one of dedicated missions. The European Space Agency's JUICE (Jupiter Icy Moons Explorer) launched in April 2023 and will enter Ganymede orbit around 2034 — the first spacecraft ever to orbit a moon other than our own. NASA's Europa Clipper launched in October 2024 and arrives at Jupiter in 2030, flying dozens of close Europa passes with ice-penetrating radar, a magnetometer, and instruments to sniff any plume material. Neither will land, but both are designed specifically to characterise the oceans and assess habitability.

Two flavours of ocean world

Not all subsurface oceans are alike. The crucial distinction is what sits at the bottom of the ocean.

• Rock-bottomed oceans. On the smaller moons — Europa and Enceladus — the ocean is thin enough that pressure at its base never reaches the threshold for high-pressure ice. The water sits directly on silicate rock. This is the configuration astrobiologists prize: water–rock contact enables hydrothermal vents, serpentinization, and the redox chemistry that powers life at Earth's mid-ocean ridges. Cassini's silica nanoparticles are the direct evidence of this at Enceladus.
• Ice-sandwiched oceans. On the large moons — Ganymede, Callisto, Titan — the ocean can be hundreds of kilometres deep, and below ~2 GPa the water freezes into dense high-pressure ice (ice VI and beyond). This forms an ice floor between the liquid and the rocky core, potentially cutting off the supply of rock-derived nutrients. Whether minerals can still diffuse or convect through that ice layer is an open and actively studied question. Some models even predict multiple stacked liquid layers between different ice phases.

Related phenomena that share the same machinery: cryovolcanism (eruption of liquid water or briny slush instead of molten rock — Enceladus's plumes are the clearest case, with Titan a strong candidate); chaos terrain and double ridges on Europa, surface expressions of liquid or warm ductile ice moving below; and the magnetic induction response itself, which is the same physics that lets geophysicists probe Earth's conductive interior with magnetotelluric soundings.

Common misconceptions and subtleties

• "The ocean is warm and tropical." No. The water hovers near its (depressed) freezing point — roughly 0 °C at the ice interface, perhaps a little above. It is liquid, not balmy. "Habitable" here means liquid water plus chemical energy, not comfortable temperatures.
• "Sunlight keeps it from freezing." Sunlight is negligible. The ocean is sustained entirely by tidal and radiogenic heat from within, with the ice shell acting as insulation. This is precisely what makes these worlds independent of the conventional surface habitable zone.
• "Tidal heating is from the planet's tides sloshing the ocean." The dominant dissipation is the periodic deformation of solid material (rocky mantle and ductile ice) as the body is flexed, not ocean-current friction. The ocean matters more as a decoupler that lets the shell flex freely than as the seat of the dissipation.
• "Eccentricity should have damped away by now." It would, if not for orbital resonances continually re-pumping it. Remove the Laplace resonance and Europa's ocean budget collapses. Resonance is not a detail — it is the reason these oceans still exist.
• "Finding an ocean means finding life." An ocean supplies one of life's requirements. Whether there is also a sufficient, sustained chemical disequilibrium (a redox gradient to eat) and direct rock–water contact remains unproven — which is exactly what Europa Clipper and JUICE are built to test.
• "All icy moons have oceans." Many do not. The body must be large enough to retain heat and, in most cases, must have a maintained source of eccentricity. Mimas, smaller than Enceladus and once thought too cold, surprised researchers in 2024 with dynamical evidence of a young ocean — but plenty of mid-size icy moons appear frozen through.