Imagine sunrise on a frozen plane. Shadows withdraw as the Sun climbs above distant mountains that rise from below the horizon. The ice sheet itself is largely featureless, with a difference in elevation of only some tens of metres over distances of many tens of kilometres — nothing slows the shadows' retreat. The sky overhead remains black, and it stays chilly, only about 35 degrees above absolute zero. This is morning on Sputnik Planum, Pluto.
The fly-by of Pluto (and its satellites) by NASA's New Horizons space probe in July 2015 revealed a spectacular world unlike any yet seen1. The informally named Sputnik Planum is of special interest, a bright, flat-floored basin around 1,200 km in diameter that is filled mainly with nitrogen ice (Fig. 1). High-resolution images1 show a surface separated into polygonal cells 10–40 km in diameter, pockmarked by pits and fed by flowing nitrogen glaciers from the surrounding highlands. In this issue, Trowbridge et al.2 (page 79) and McKinnon et al.3 (page 82) investigate this polygonal terrain and conclude that it is continually and quickly resurfaced by convection, making it one of the youngest surfaces in the Solar System. Pluto therefore joins Europa, Enceladus, Titan and Triton as a small and icy but geologically dynamic body of the outer Solar System — a far cry from the cold, dead worlds one might expect so far from the Sun.
The nitrogen ice identified by New Horizons is a structurally weak solid with a very low melting point (63 kelvin), and should flow viscously even at Pluto's low temperatures4. As in other planetary bodies, Pluto's interior is warmer than its surface because it is heated by the decay of long-lived radiogenic isotopes in the rocky component. How this heat escapes through Sputnik Planum has consequences for its surface geology. For a layer of weak nitrogen ice at least 0.5–1 km thick, the most efficient heat-transfer mechanism is convection.
Because the material is heated from the bottom, the heat will cause localized thermal expansion, making the heated material less dense than the rest of the overlying ice. In convection, the less-dense material is buoyant and will rise, carrying its heat content towards the surface, where it cools and then sinks. Viscous drag resists this buoyancy-driven movement, and convection can occur only if the buoyancy overwhelms the viscous resistance.
This competition can be quantified. The ratio of buoyant to viscous forces in a layer defines a dimensionless parameter known as the Rayleigh number. If the Rayleigh number is greater than a critical value, then the material convects. Both Trowbridge et al. and McKinnon et al. found that the Rayleigh number of the polygonal terrain is several orders of magnitude greater than the critical value. Their results indicate that the nitrogen ice is vigorously convecting and that the cellular patterns are the tops of convection cells.
In addition, both groups report that the convective flow speeds are in the range of centimetres per year, meaning that the surface turns over in about 500,000 to 1 million years. This rapid resurfacing explains the lack of impact craters on the ice sheet. (In general, the older a planetary surface, the more impact craters will have formed.)
Although the two papers report the same primary result, they differ in their conclusions about the convective regime, which determines the width-to-depth aspect ratio of the convection cells and hence the thickness of the nitrogen-ice layer. Trowbridge et al. argue that variations in the viscosity of the ice due to differences in stress and temperature across the layer are small enough that convection occurs in the Rayleigh–Bénard regime, which is characterized by the formation of cells that have widths similar to their depths5. The cell size of 10–40 km thus implies a layer thickness of at least 10 km.
By contrast, McKinnon et al. argue that the temperature dependence of the nitrogen ice causes 'sluggish lid' convection, in which the viscosity is higher at the cooler surface than in the interior5. As the name suggests, this yields a slower-moving surface layer and cells that are much wider than they are deep, making the depth of the layer 3–6 km. The authors support this conclusion with numerical modelling that reproduces convection cells with sizes and surface topography that are consistent with observations.
The layer thickness has important implications for Pluto's geological history. On the basis of the shape and ellipticity of the basin that holds Sputnik Planum, McKinnon and colleagues note that it is most probably an ancient impact crater3. From scaling of other examples in the Solar System, it is known that an impact basin of this size can easily accommodate the depth of nitrogen ice estimated by McKinnon et al., but not the depth estimated by Trowbridge and colleagues. Their deeper prediction requires a more complicated explanation of basin formation and evolution. Perhaps the weight of the nitrogen ice caused the basin to subside, for example.
Both papers report that the quantity of ice in the basin is equivalent to a global layer several hundred metres in depth, commensurate with Pluto's total budget of nitrogen. But neither satisfactorily addresses how so much of the nitrogen budget could have collected there — was it for climatological reasons, as Trowbridge and co-workers speculate, or for glaciological reasons, as McKinnon et al. suggest? Clearly, this localization of nitrogen was a major event in Pluto's evolution that needs to be explored. Fortunately, New Horizons continues to transmit data from its Pluto encounter back to Earth. It is to be hoped that these two papers will be the first step towards a deeper understanding of this distant world.