Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Planetary Science. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 17 January 2021

The Structure and Dynamics of the Atmospheres of Pluto and Tritonfree

  • Angela M. ZaluchaAngela M. ZaluchaA.M. Zalucha Consulting
  •  and Jason CookJason CookPinhead Institute and SETI Institute


In addition to ground-based observations beginning in the 1970s, NASA’s Voyager 2 spacecraft flew by Triton in 1989, and NASA’s New Horizons spacecraft flew by Pluto in 2015. Prior to the flyby of New Horizons, Pluto and Triton were termed “sister worlds” due to what appeared to be a high degree of similarity in solid-body density, surface ices, diameter, and surface pressures. Despite being small, cold, icy bodies, both Pluto and Triton have been found to have atmospheres that behave as a continuous fluid up to 300 km altitude above the surface and thereby have a defined temperature, surface pressure, and global general circulation (wind). The primary constituent of these atmospheres is molecular nitrogen, with methane and carbon monoxide comprising the largest abundances of trace gases. The surface pressure as measured in the 2010s on both worlds is of the order of 10 microbars (1 Pa = 10 µbar), for these exotic atmospheres exchange mass between sublimation of surface ice and deposition of nitrogen over the course of each body’s year. Ground-based stellar occultation measurements observed a dramatic change in surface pressure, which one study found was as much as a factor of two increase between 1988 and 2003 on Pluto, presumably due to Pluto’s seasonal volatile cycle. Voyager 2 observed plumes and surface “streaks” on Triton, while New Horizons observed dunes (indicating wind speeds of 1–10 m s1) as well as streaks, evidently indicating the presence of surface and near-surface winds.

While wind velocity aloft has not been directly measured on Pluto or Triton, 3-D general circulation modeling studies of both worlds have shown zonal (east–west) wind speeds of the order of 10 m/s, meridional (north–south) wind speeds of the order of 1 m/s, and extremely weak vertical wind speeds.

In 2015, New Horizons showed that Pluto and Triton were much more different than previously thought. New Horizons uncovered many spectacular views of Pluto’s atmosphere. First, while hydrocarbon haze was observed on Triton, Pluto had multiple, very distinct stratified haze layers bearing a similar appearance to the layers of an onion. Second, Pluto’s surface elevation was found to be largely inhomogeneous (in contrast to Triton) in the form of a large depression (Sputnik Planitia). Third, the characteristics of the surface markings on Pluto were found to be different than the streaks observed on Triton, which has implications for surface wind patterns.

Further major discoveries made by New Horizons included evidence for many hydrocarbon species in trace concentrations, a lower than expected surface pressure, which could previously only be indirectly ascertained from ground-based observations, and a higher mixing ratio of methane at higher altitudes than at lower due to gravitational diffusive separation. Using radio occultation experiments (not conducted by Voyager 2 at Triton), New Horizons confirmed the existence of a stratosphere (temperature increasing with height) extending to 25 km altitude at both the entry and exit locations. The entry location had a shallow troposphere (temperature decreasing with height) extending to 3.5 km altitude above the surface, while the exit location did not.

Pluto and Triton’s Atmospheres: A Class of Their Own

Pluto and Triton belong to a unique class of solar system objects because their atmospheres are unlike any of the terrestrial atmospheres (including Titan) or gas giant planets. The atmospheres of Pluto and Triton constitute their own class: cold atmospheres (as of the 2010s) at 10–20 µbars surface pressure (1 Pa = 10 µbar), a circulation dominated by condensation flow, a transparent atmosphere, and a chemical composition composed primarily of N2 in addition to CO, CH4, and trace amounts of other hydrocarbons.

Both Pluto and Triton have been visited by spacecraft only a single time each—NASA’s New Horizons in 2015 (Stern et al., 2015) and NASA’s Voyager 2 in 1989 (Broadfoot et al., 1989), respectively. Both missions were highly successful in their respective flybys of these worlds, resulting in stunning imagery and unprecidented measurements of temperature, pressure, and chemical composition. Triton was found to have rising plumes of dark material, clouds, and surface streaks potentially caused by aeolian features. Pluto was found to have a stratified haze layer and a highly heterogenous surface ice distribution (which has implications for the distribution of atmospheric heating and temperature).

The tremendous distances of Pluto and Triton from Earth (30 to 50 AU) relative to other bodies in the solar system mean that spacecraft missions to these bodies are costly, both in terms of the time it takes for a spacecraft to reach them and the ensuing financial costs. Both before and after the respective flybys of the New Horizons and Voyager 2 spacecraft, Earth-based observers and theoretical modelers have played just as crucial a role as the close-up observations by these spacecraft. By far the Earth-based techniques that have yielded the most insight about the atmosphere are spectral measurements and observations of stellar occultation events. A stellar occultation is an astronomical event that occurs when a foreground object in the solar system passes in front of a background star for an Earth-based observer (see Figure 1 for a schematic).

Figure 1. Diagram of differential refraction during a stellar occultation. Light from the star (assumed to be to the left of the diagram) is initially parallel and is denoted by the black arrows. The solid body is given in brown, and its atmosphere (relative size greatly exaggerated) in blue. Upon encountering the atmosphere, the light rays are bent according to the refractive properties of the atmosphere, which depend on density. Since atmospheric density increases as one moves closer to the solid body’s center, light rays near the surface are bent at a larger angle. This differential refraction causes the light rays to spread out and appear as an attenuated signal. The observer, in the shadow plane, sees light of varying intensity, referred to as the light curve. From the light curve, properties of the atmosphere, including temperature, pressure, density, composition, wave activity, and composition, may be inferred by a variety of methods (see Elliot & Olkin, 1996; Elliot, Person, & Qu, 2003, for a detailed review of occultation analysis methods).

Source: A. M. Zalucha.

Modeling efforts have also played a key role in understanding Pluto’s and Triton’s atmospheres. Early models predicted the temperature as a function of height given abundances of gaseous species (Hubbard, Yelle, & Lunine, 1990; Strobel, Zhu, Summers, & Stevens, 1996; Yelle & Lunine, 1989). Others have predicted the transport of volatile materials (and subsequently surface pressure, surface temperature, and surface ice distribution) as a function of time (Hansen & Paige, 1992, 1996; Young, 2012, 2013). In the 2010s, state-of-the-art general or global circulation models emerged for Pluto and Triton that simultaneously predict 3-D temperature, 3-D wind speed (which cannot be observed remotely), surface pressure, and surface ice distribution as a function of time (Bertrand & Forget, 2016, 2017; Forget et al., 2017; Toigo et al., 2015; Zalucha & Gulbis, 2012; Zalucha & Michaels, 2013).

Uncovering the Mysteries of the Atmospheres of Distant Icy Worlds: Stellar Occultations, Spectra, and Early Models of Pluto

Stellar Occultations

Brosch (1995) observed a stellar occultation by Pluto on August 19, 1985 from Wise Observatory in Israel, but due to the poor observing conditions, at the time the presence of an atmosphere was disputed. Later, authors concluded that the primary atmospheric constituent was “probably” N2, CO, or CH4.

The first high-quality stellar occultation by Pluto was observed on June 9, 1988 at several locations in Australia and by the Kuiper Airborne Observatory (KAO) (Elliot et al., 1989; Hubbard, Hunten, Dieters, Hill, & Watson, 1988; Millis et al., 1993). This event is considered the first conclusive detection of the atmosphere of Pluto. Model fits to the KAO light curve (Figure 2), which had the highest signal-to-noise ratio (SNR), assuming a pure CH4 atmosphere, derived a temperature of 67 ± 6 K at an altitude of 1,214 ± 20 km. A puzzling feature, even to this day (2019), in the KAO light curve is the physical cause of a kink at the 40% normalized intensity (flux) level, at which point the light curve flux suddenly starts decreasing more sharply with time. Light curve data are ambiguous in that the flux may diminish either due to extinction of the starlight by an optically thick constituent (such as aerosols or liquid or ice clouds) in the atmosphere or due to a change in the refractivity (and hence spreading of the light rays) caused by a change in the temperature gradient. This topic is revisited throughout this article.

Figure 2. Light curve from a stellar occultation by Pluto observed by the KAO on June 9, 1988 (Elliot et al., 1989). Note the “kinks” at about 55 and 135 seconds, where the slope of the curve changes abruptly.

Source: A. M. Zalucha.

Elliot and Young (1992) developed an analysis method to model occultation events by small body atmospheres, a variant of which is still in use in 2019. The temperature of the atmosphere is assumed to depend on the radius from the body’s center according to a power law. Inspired by the kink in the 1988 KAO light curve, Elliot and Young (1992) included free parameters to describe a haze layer (haze scale height, haze opacity at a reference altitude, and depth of haze layer), where the haze opacity decreased exponentially with height from the surface. Using this prescription, the kink could be explained mathematically, though not chemically.

A second set of occultations was observed 14 years later on July 20, 2002 and August 21, 2002 from locations in South America and the western United States (including Hawaii), respectively (Elliot, Ates, et al., 2003; Pasachoff et al., 2005; Sicardy et al., 2003), which resulted in more questions than answers. First, the kink seen in the 1988 KAO light curve had disappeared; second, the analysis using the Elliot and Young (1992) model implied a factor of two increase in the surface pressure (though the reanalysis of Zalucha, Gulbis, Zhu, Strobel, & Elliot [2011], using a physical model of Pluto’s atmosphere, indicated the increase was less than a factor of two). Moreover, Elliot, Ates, et al. (2003) noted a wavelength dependence in the August 21, 2002 occultation data that was too large to be due to the dependence of N2 refractivity on wavelength, and was thus attributed to a wavelength-dependent haze extinction, furthering the argument for a haze layer on Pluto in the era prior to the New Horizons flyby.

From a subsequent stellar occultation observed on June 12, 2006 at various sites in Australia by Elliot et al. (2007) and by E. F. Young et al. (2008) (Figure 3), the analyses showed only subtle changes in atmospheric structure from the 2002 observations.

Figure 3. Light curve from a stellar occultation by Pluto observed by at the Anglo-Australian telescope on June 12, 2006 (E. F. Young et al., 2008), an example of a more typical and high SNR light curve.

Source: A. M. Zalucha.

Another stellar occultation on March 18, 2007 observed in Hawaii by Person et al. (2008) showed that the atmosphere was essentially unchanged from the 2006 occultation. Meanwhile, Olkin et al. (2009) showed a 28% increase in pressure at a reference radius of 1,275 km between their observations of the July 31, 2007 occultation and the June 12, 2006 occultation (observed by E. F. Young et al., 2008). Adding further to the haze debate, L. Young et al. (2008) observed no wavelength dependence in their observations of the July 31, 2007 occultation at 0.5 and 0.8 µm.

The light curve obtained on March 18, 2007 at the Multiple Mirror Telescope Observatory (MMT) had a very high SNR and contained variations during immersion and emersion that were identified as waves in Pluto’s atmosphere. What is remarkable is that each half of the light curve was nearly identical, indicating coherent wave structures between both limbs of Pluto’s disk visible at the time of the occultation.

Person et al. (2008) (incorrectly) interpreted the waves as Rossby waves and calculated an upper limit for the zonal wind speed of 3 m s1 based on the critical velocity required for vertical propagation. However, Hubbard et al. (2009) provided evidence that the waves are instead gravity waves.

While observations of stellar occultation events are usually quite rare (only one having been observed between 1988 and 2002), 14 were observed in the following 13 years (2003–2015). The relatively large number of occultation events was due to Pluto passing through the galactic plane, where there are more possibilities for it to occult stars (and no doubt excitement about the then upcoming New Horizons encounter provided additional motivation for prediction and observation). Elliot et al. (2007) and E. F. Young et al. (2008) observed an occultation on June 12, 2006 and found that the atmospheric surface pressure had not increased since the 2006 occultation observation. Olkin et al. (2014) observed an occultation on July 31, 2007, also showing no increase in pressure and, for the first time, observing a central flash (an effect created by diffraction when the occulting body’s exact center crosses in front of the occulted star). From analysis of the central flash data, these authors concluded that a thermal gradient model solution, rather than a haze only or combination of haze and thermal gradient, best explained the data, thus arguing against haze on Pluto in 2007.

Further occultations were observed on August 25, 2008 (Buie et al., 2009), February 14, 2010 (L. Young et al., 2010), July 4, 2010 (Pasachoff et al., 2010; Person et al., 2010; L. Young et al., 2010), May 4, 2013, and Universal Time (UT) September 9, 2012 (Bosh et al., 2015), July 31, 2014 (Person et al., 2014), and three in July 2014 (Pasachoff et al., 2015) that showed the surface pressure continuing to be stable (a relief, as models had shown it to be possible for the atmosphere to be non-existent by the time of the New Horizons encounter). An event observed by Person et al. (2013) on June 23, 2011 also showed that the surface pressure continued to be stable, but resulted in a light curve that had a kink shape like the June 9, 1988 occultation. These authors concluded that the lower atmosphere evolves on a faster timescale than the upper atmosphere (the radius from Pluto’s center greater than 1,290 km). Dias-Oliveira et al. (2015) obtained 12 light curves from occultations on July 18, 2012 and May 4, 2013 that they simultaneously analyzed; they found that the then currently measured CO abundance was too small to explain the drop in temperature with height in the mesosphere (1,215–1,390 km from Pluto’s center). They proposed HCN as an additional coolant, but found that it would need to be present in amounts greater than expected from condensation arguments.

Finally, Pasachoff et al. (2017) observed an occultation on June 29, 2015, weeks before the New Horizons encounter. This observation was important because again a central flash was observed, evidence for atmospheric layering was seen in the light curve (spikes), and ground-truth was provided to normalize the occultation data to the soon to be obtained New Horizons spacecraft data.


Image tube spectra taken by Benner, Fink, and Cromwell (1978) in the range of 6,800 to 9,000 Å of Pluto potentially indicated gaseous CH4 on Pluto in the 8,900 Å band, but because the line was so near the edge of the wavelength range sampled, its validity was inconclusive. A few years later, low-resolution spectrograph and CCD array measurements taken by Fink et al. (1980) in the 5,800–16,000 Å wavelength range showed that the 8,900 Å band was saturated compared to other bands, indicating atmospheric CH4. However, their estimate of the CH4 (partial) pressure 15 Pa was of the order of a factor of 1,000 too high. Buie and Fink (1987) performed absolute spectrophotometry of Pluto in the wavelength range of 5,600 to 10,500 Å over four nights and found that their observations could be explained by either an anisotropic surface distribution of CH4 frost and a clear layer of CH4 gas or no gaseous CH4 whatsoever. For the model with the CH4 present in the atmosphere, the upper limit on the one-way column abundance was revised down to 5 m-am (compared with the absolute value of 27±7 m-am of Fink et al., 1980), which agrees with later ground-based and New Horizons measurements (as does the non-homogeneous distribution of surface CH4). Stansberry, Lunine, and Tomasko (1989) investigated possible mechanisms for haze production, but found they needed CH4 dissociation rates and absorption strengths to be beyond reasonable limits (based on analogs with other hazes in the outer solar system like Titan and Triton).

Subsequent measurements by Owen et al. (1993) of solar reflection spectra in the near IR revealed the presence of surface ice composed of CO, N2, and CH4, where the N2 ice was more abundant than the other two by a factor of 50. Furthermore, from vapor-ice equilibrium calculations, it was determined that N2 was the dominant atmospheric gas in terms of partial pressure by several orders of magnitude compared with the other two gases. If the ices were in vapor pressure equilibrium with the surface gas, the relatively large abundance of N2 ice implied that the atmosphere must also be predominately N2.

Continued spectral measurements improved constraints on the composition of Pluto’s atmosphere. Young, Elliot, Tokunaga, de Bergh, and Owen (1997) used an echelle spectrograph at NASA’s IRTF facility on Mauna Kea, Hawaii on May 25–26, 1992 with a spectral resolution of 13,300. The spectral range was 1,661.8–1,666.9 nm (often referred to as simply the 1.7 µm line). Because of the superior resolution to previous measurements, the gaseous lines (specifically, the 2ν‎3 band) were able to be detected directly rather than inferred from surface measurements. These authors derived a column height value of 3.222.34+8.46×1019 molecule cm−2, which, using vapor pressure arguments from the results of Owen et al. (1993), corresponded to a partial pressure of 0.0720.052+0.189 µbar.

Lellouch et al. (2009) combined spectral observations of the 2ν‎3 band covering the 1,642–1,650, 1,652–1,659, 1,662–1,670, and 1,672–1,680 nm ranges (mean spectral resolution of 60,000) on August 1, 2008 at the European Southern Observatory Very Large Telescope UT1 (Antu) 8.2 m telescope and constraints obtained from the June 12, 2006 occultation (E. F. Young et al., 2008). These authors inferred that Pluto’s surface pressure was in the range of 6–24 µbar, the methane mixing ratio was 0.5±0.1%, and the maximum troposphere depth, if present, was 17 km. Lellouch, de Bergh, Sicardy, Käufl, and Smette (2011) followed up on these results with additional spectral measurements on July 27 and 29, 2010 (this time also encompassing a fraction of the ν‎1 + ν‎4 and ν‎3 + ν‎4 bands of CH4 at 2,312–2,325, 2,330–2,342, 2,345–2,356, and 2,359–2,370 nm ranges), in which they obtained mixing ratios of 0.6+0.6−0.3% and 0.5+10.25× 103 for CH4 and CO (J=2–0 line), respectively. Greaves, Helling, and Friberg (2011) also confirmed the detection of atmospheric CO from observations of the 1.3-mm wavelength J=2–1 line taken at the James Clerk Maxwell Telescope (JCMT) on Mauna Kea, Hawaii, observing over three nights in August 2009 and eight nights in April–May 2010. Furthermore, Greaves et al. (2011) detected an extended source of CO; however, this observation was neither confirmed nor refuted by New Horizons.

Lellouch et al. (2017) observed Pluto shortly before the New Horizons flyby using the ALMA interferometer on June 12–13, 2015, and detected the CO (J=3–2) and HCN (J=4–3) rotational transitions (including the hyperfine structure of HCN), thus providing a strong confirmation of the presence of CO and the first observation of HCN in Pluto’s atmosphere. The CO and HCN lines probed Pluto’s atmosphere up to 450 km and 900 km altitude, respectively, with a large contribution due to limb emission. The CO detection yielded a CO mole fraction of 515±40 ppm (for a 12 µbar surface pressure) and clear evidence for a well-marked temperature decrease above 30–50 km altitude and a best-determined temperature of 70±2 K at 300 km. The line shape of the observed HCN resulted in a derived mole fraction <i» 1.5 × 105 above 450 km and a value of 4 × 105 near 800 km. HCN was also present in the bottom 100 km of the atmosphere, with a 108 to 107 mole fraction, depending on the precise stratopause temperature.


Due to the computational expense of 2- or 3-D models, initially models of Pluto and Triton were carried out with one spatial dimension and time (either implicitly or explicitly, depending on the model). One category of models exploits the surface-atmosphere mass exchange to determine properties of Pluto’s (and Triton’s) atmosphere. Relatively recently after the discovery of Pluto’s atmosphere, Hansen and Paige (1996) published a study about the distribution of ice on Pluto’s surface. The model included one atmospheric layer such that the temperature of the atmosphere and surface pressure could be predicted at a given time and latitude (there was no longitudinal degree of freedom) as well as mass transport between latitudes.

Given the atmospheric simplicity, this model could be executed for many Pluto years until a repeatable seasonal surface ice cycle was reached. The study was conducted as a parameter sweep over surface thermal inertia, the albedo of the ice-free surface, ice albedo and emissivity, and total N2 inventory (the total amount of N2 contained in the ice and the atmosphere). While the study could not constrain any of these parameters, it predicted distributions of ice in space and time (as compared with Mars, which also has a volatile cycle), which is a direct consequence of Pluto’s eccentric orbit and high obliquity.

Young (2013) performed a similar study to Hansen and Paige (1996), but with the advantages that there was then more data of Pluto’s changing surface pressure with time and surface thermal properties and that a wider parameter space could be explored with the increase in processing speed of more modern computers. Upon filtering the results to those that were consistent with observational data, three main categories of equilibrium states were found: Permanent Northern Volatile (permanent ice cap at the north pole), Exchange with Pressure Plateau (two volatile caps for a long period after the perihelion equinox), and Exchange with Early Collapse (northern volatiles lost shortly after perihelion). As was found by Hansen and Paige (1996), the latter category describes a case where Pluto’s atmosphere is completely lost to surface ice for some period of the year. Bertrand and Forget (2016) used a volatile transport model that included three volatile species (N2, CH4, and CO) to show that the Sputnik

Planitia depression on Pluto’s surface could be of atmospheric origin, as their model predicted the accumulation of N2 in the basin over thousands of years. A follow-up paper by Bertrand et al. (2018), which integrated the model over millions of years, showed that (1) obliquity dominates the N2 cycle and that over one obliquity cycle, the latitudes of Sputnik Planitia between 25° S and 30° N were dominated by N2 condensation, while the northern regions between 30° and 50° N were dominated by N2 sublimation; (2) a net amount of 1 km of ice had sublimed at the northern edge of Sputnik Planitia during the last 2 million years; (3) N2 ice is not stable at the poles but rather in the equatorial regions, in particular in depressions, where thick deposits may persist over tens of millions of years, before being trapped in Sputnik Planitia; and (4) the minimum and maximum surface pressures obtained over the simulated millions of years remain in the range of milli-Pascals and Pascals, respectively.

Another category of models determines temperature in a vertical column from energy balance arguments. From the new constraints on Pluto’s atmosphere from the June 9, 1988 occultation event, Yelle and Lunine (1989) developed a non-local thermodynamic equilibrium (non-LTE) radiative-conductive model that included the effects of heating by the CH4 3.3 µm band, cooling by the CH4 7.6 µm band, and thermal conduction. They argued that Pluto’s atmosphere must contain some molecule or molecules heavier than CH4 to be consistent with the scale height of 59.7 ± 1.5 km obtained by Elliot et al. (1989). Yelle and Lunine (1989) suggested CO, N2, and Ar as the most suitable candidates. Hubbard et al. (1990) used a model similar to Yelle and Lunine (1989) to show that if the atmospheric temperature profile exhibited a strong temperature inversion near the surface and an isothermal atmosphere above, a kink could be produced, as observed in the 1988 KAO light curve (later revised by Lellouch, 1994). Stansberry, Lunine, Hubbard, Yelle, and Hunten (1994) used an idealized temperature profile, which they adjusted to fit the properties of the 1988 light curve, and found that the temperature gradient in the inversion region must be 30 K km1 to reproduce the kink. Strobel et al. (1996) presented a much more sophisticated radiative-conductive model by adding CH4 heating in the 2.3 µm band and CO cooling (due to numerous rotational lines), as well as the effects of vibrational energy transfer in the CH4 molecule, but found they were not able to achieve the steep stratospheric gradient required by Stansberry et al. (1994).

In a novel approach to ascertaining atmospheric parameters from light curves, Zalucha, Gulbis, et al. (2011) and Zalucha, Zhu, Gulbis, Strobel, and Elliot (2011) used an updated version of the Strobel et al. (1996) physical model of Pluto’s atmosphere to calculate model light curves using a forward modeling approach. These authors found surface pressures (depending on the light curve being analyzed) that ranged from 8.1 to 13.9 µbar CH4 mixing ratios in the range of 0.18 to 0.94%. Zalucha, Zhu, et al. (2011) noted that if a troposphere was present, it must be less than 1 km in depth.

While the Strobel et al. (1996) model was very detailed in its calculation of temperature, it is a 1-D model and did not take into account heating in the atmospheric column by advection or latent heating. In 2011, 3-D models of Pluto and Triton from four separate research groups exploded onto the conference scene of a specific type: general (or global) circulation models (GCMs) (Michaels & Young, 2011; Miller, Chanover, Murphy, & Zalucha, 2011; Vangvichith & Forget, 2011; Zalucha & Gulbis, 2011) (in addition to a model previously presented by Mueller-Wodarg, Yelle, Mendillo, & Aylward [2001] many years earlier). GCMs are a type of geophysical fluid model that integrates the Navier-Stokes equations forward in time over long (climate) time scales on a sphere. They often include additional external terms such as radiative heating and cooling; surface-atmosphere mass, momentum, or heat exchange; and chemical cycles. These models are global in extent and simultaneously predict the state of the atmosphere in a consistent way. They are particularly useful in studies of planetary atmospheres, where measurements, especially those of wind, are scarce.

Zalucha and Gulbis (2012) were the first to publish a Pluto GCM (PGCM). Their PGCM was simplified in the sense that (1) it was in 2-D (latitude, height, and time), and (2) it did not include a volatile cycle, which is believed to exist on Pluto given the presence of atmospheric ices and gases. A robust result was the lack of a Hadley cell-like overturning circulation. Such a circulation would likely be inhibited given the strong temperature inversion that was predicted by radiative-conductive models. Zalucha and Michaels (2013) expanded the work of Zalucha and Gulbis (2012) to 3-D, adding the longitudinal direction. Again, there was no evidence of a Hadley cell or overturning circulation.

Toigo et al. (2015) employed a novel approach with their PGCM by running it in two phases. First, they integrated over time a 2-D model similar to the volatile transport models with a simplified atmosphere, so as to quickly equilibrate the surface ice. Then, a full 3-D atmospheric PGCM was evoked, which included the volatile cycle and used an updated radiative-conductive model from Zhu, Strobel, and Erwin (2014) for atmospheric heating and cooling. The Zhu et al. (2014) model is the successor to Strobel et al. (1996) as it added solar far ultraviolet and extreme ultraviolet heating in the upper atmosphere and adiabatic cooling due to hydrodynamic expansion (including atmospheric escape). The main result from the study of Toigo et al. (2015) was that near-surface winds generally follow a sublimation flow from the sunlit polar cap to the polar night cap, with a Coriolis turning of the wind as the air travels from pole to pole. They demonstrated the strong contribution of nitrogen sublimation and deposition to Pluto’s atmospheric circulation, in contrast to Mars, where the sublimation flow is a minor component to the total atmospheric circulation (Haberle et al., 1993).

With the new information gained by New Horizons, a new generation of PGCMs was born. In particular, the unexpectedly heterogeneous surface topography, including the large depression named Sputnik Planitia, has been shown to play a greater role in Pluto’s general circulation than previously known (Forget et al., 2017; Soto, Rafkin, & Michaels, 2016). In the reference simulation with surface N2 ice exclusively present in Sputnik Planitia, the global circulation was only forced by radiative heating gradients and remained relatively weak. Surface winds were locally induced by topographic slopes and by N2 condensation and sublimation around Sputnik Planitia. However, the circulation could be more intense depending on the exact distribution of surface N2 frost.

Under certain conditions, a global condensation flow was created, inducing strong surface winds everywhere. The Forget et al. (2017) model allowed N2, CH4, and CO condensation. These authors found that N2 and CO did not condense in the atmosphere, but CH4 ice clouds could form during daytime at a low altitude near the regions covered by N2 ice (assuming that nucleation is efficient enough). Bertrand and Forget (2017) used the same GCM to help model haze photochemistry, but the discussion is beyond the scope of this article.

Recent observations by New Horizons (Gladstone et al., 2016) showed that Pluto’s atmosphere was much colder than predicted based on pre-encounter theoretical models, suggesting an unknown cooling mechanism. Atmospheric gas molecules, particularly water vapor, in addition to CO, C2H2, and HCN (Forget et al., 2017; Gladstone et al., 2016), were proposed as a coolant (Strobel & Zhu, 2017); however, because Pluto’s thermal structure is expected to be in radiative–conductive equilibrium, the required water vapor would need to be supersaturated by many orders of magnitude under thermodynamic equilibrium conditions. Zhang, Strobel, and Imanaka (2017) added the thermal effects of haze to the radiative-conductive energy balance model of Zhu et al. (2014) and found that haze particles had substantially larger solar heating and thermal cooling rates than gas molecules. This finding makes Pluto’s atmosphere unique among solar system planetary atmospheres, as its radiative energy equilibrium is controlled primarily by haze particles instead of gas molecules.


Voyager 2 Flyby

During the Voyager 2 flyby in 1989, a variety of surface and atmospheric features were observed. Surface plumes extending to 8 km altitude (Smith et al., 1989) have sparked particular interest. Images from Voyager 2 showed two well-observed plumes, denoted as the “west” and “east” plumes, located at 50° and 57° latitude, respectively. The west plume appears as a dark column that abruptly terminates at 8 km altitude. Connected to this plume is a more diffuse cloud that extends westward for at least 150 km. Similarly, the east plume also rises to an altitude of 8 km and is directed westward at altitude. These plumes act as a tracer that indicate prevailing wind direction at this altitude. Yelle, Lunine, Pollack, and Brown (1995) interpreted the height of the plumes as the depth of a thermally defined troposphere, while Zalucha and Michaels (2013) used a Triton GCM to argue that Triton’s lower atmospheric structure is dynamically maintained.

Hansen, Terrile, McEwen, and Ingersoll (1990) identified a terminator cloud in Voyager 2 images that moved 13 m s1 eastward and was located at an altitude of 5 km. However, they point out that because the inferred velocity of this cloud is so similar to the velocity of the terminator of Triton’s surface, the cloud may in fact be a stationary, elongated east–west cloud that is being illuminated at different points along the cloud by the low-altitude sunlight. Hansen et al. (1990) also observed several crescent streaks at 1–3 km altitude and surface streaks that potentially indicate wind direction at < 1 km altitude. The direction of the crescent streaks was nearly uniformly westward, while the surface streaks were variable but mainly northeastward. Voyager 2 measured an atmospheric pressure of 14 µbar (Broadfoot et al., 1989) and a CH4 partial pressure of 2.45 nbar (Herbert & Sandel, 1991; Strobel & Summers, 1995).

Stellar Occultations

From a stellar occultation that occurred on November 4, 1997, Elliot et al. (1998) observed that the pressure aloft had increased significantly since the Voyager 2 encounter and measurements from a 1995 stellar occultation. They concluded that the surface pressure of Triton doubles every 10 years during the 1989–1997 period of observation. Using spectroscopic measurements, Lellouch, de Bergh, Sicardy, Ferron, and Käufl (2010) found that in 2009 the partial pressure of CH4 had increased to 9.8 ± 3.7 nbars (a factor of 4 since Voyager 2). A handful of other stellar occultations by Triton have been observed: July 10, 1993 (Olkin et al., 1997), August 14, 1995 (Olkin et al., 1997), and May 21, 2008 (Sicardy et al., 2008).

Unfortunately, as Triton started to pass through a fairly sparse star field, occultation observations have been much more difficult to make.


There are few studies of Triton prior to the Voyager 2 flyby in 1989; studies by Cruikshank and Silvaggio (1979) and Apt, Carleton, and Mackay (1983) revealed CH4 bands. During the Voyager 2 encounter, N2 was detected spectroscopically in the atmosphere by Broadfoot et al. (1989), who also identified it as the primary constituent. CH4 was also confirmed as an atmospheric constituent. Lellouch et al. (2010) observed CH4 and CO using Earth-based spectroscopy. It is also possible that Ar is present in the atmosphere at up to 10% abundance, but it is nearly impossible to detect it spectroscopically (McKinnon, Lunine, & Banfield, 1995). Near IR observations taken in 2010, 2011, and 2013 by Merlin, Lellouch, Quirico, and Schmitt (2018) suggest that the atmospheric pressure over 2010–2013 was similar to that during the Voyager 1989 epoch.


Ingersoll (1990) and Yelle et al. (1995) calculated the speed of the volatile (condensation) flow necessary to keep the summer and winter hemispheres at the same temperature, assuming both hemispheres were uniformly covered with surface ice. In order to estimate the depth of the flow, they appealed to classical Ekman layer (the near-surface layer of the atmosphere where there is balance between the pressure gradient force, the Coriolis force, and surface frictional forces) theory and derived a meridional flow speed of 5 m s1. They also predicted an anticyclone (flow in the opposite direction of the solid body’s rotation) at the south pole. The vertical 1-D model of Yelle et al. (1991) assumed that the energy balance in Triton’s lower atmosphere was given by the sum of radiative heating by CH4 at 3.3 µm, radiative cooling by C2H2 at 13.8 µm, vertical conduction, vertical turbulent mixing, and release of latent heat due to condensation. This model predicted a troposphere depth of about 15 km, which was within a factor of 2 of the observed plume heights. The predecessor to the Hansen and Paige (1996) model for Pluto’s surface ice, Hansen and Paige (1992), predicted the surface ice distribution as a function of latitude and time. Because Triton’s surface properties were not well known, a wide variety of plausible scenarios were obtained. Hansen and Paige (1992) were unable to find a set of parameters that provided a robust match between the modeled albedo boundaries and the albedo boundaries observed during the Voyager 2 flyby.

The development of Triton GCMs (Miller et al., 2011; Mueller-Wodarg et al., 2001; Vangvichith & Forget, 2011) occurred simultaneously with PGCMs, because at the time, it was thought that the two atmospheres were very similar. Zalucha and Michaels (2013) found with their Triton GCM that, as in the PGCM results, the meridional temperature gradient in the Triton GCM results was weak. The winds in the frictional boundary layer, which extended to 8 km, were weak. There was practically no vertical motion and the global surface pressure variation was small compared to the globally averaged value (in agreement with Yelle et al., 1995). Note that the vertical rising motion observed within Triton’s surface plumes is not inconsistent with this Triton GCM result for a number of reasons. First, if the plume material is composed of a different material or is warmer than the background atmosphere, or both, this would allow the plumes to be buoyant. Second, the plumes may move vertically if they have a nonzero upward initial velocity (see Kirk, Soderblom, Brown, Kieffer, & Kargel [1995] for a detailed discussion of plume sources). Third, the Triton GCM resolution is much coarser than the horizontal area of the plumes. The GCM may have been averaging out any upward motion from the plumes with downward motion elsewhere in the grid box, resulting in zero net velocity.

Comparative planetary science between Pluto and Triton could be revisited in the wake of updated models and the new information from New Horizons. Strobel and Zhu (2017) updated the model of Zhu et al. (2014) to include Voigt line profiles in the Pluto radiation module with the latest spectral database and adapted the Zhu et al. (2014) model to Triton’s atmosphere by including additional parameterized heating due to the magnetospheric electron transport and energy deposition. According to Strobel and Zhu (2017), the CH4 mixing ratio profiles played central roles in differentiating the atmospheres of Pluto and Triton. On Pluto, the surface CH4 mole fraction was in the range of 0.3–0.8%, sufficiently high to ensure that it was well mixed in the lower atmosphere and not subject to photochemical destruction. Near Pluto’s exobase, CH4 attained comparable density to N2 due to gravitational diffusive separation and escaped at 500 times the N2 rate (1023 N2 s1). In Triton’s atmosphere, the gaseous CH4 mole fraction at the surface was of the order of 0.015%, sufficiently low enough to ensure that it was photochemically destroyed irreversibly in the lower atmosphere and that N2 remained the major species, even at the exobase. With solar EUV power only, Triton’s upper thermosphere was too cold and magnetospheric heating, approximately comparable to the solar EUV power, was needed to bring the N2 tangential column number density in the 500–800 km range up to values derived from the Voyager 2 UVS observations (Broadfoot et al., 1989). Due to their cold exobase temperatures relative to the gravitational potential energy wells that N2 resides in, atmospheric escape from Triton and Pluto was not dominated by N2 Jeans escape but by CH4 from Pluto and H, C, N, and H2 from Triton (Strobel & Zhu, 2017).

A New Era of Pluto and Triton Atmospheric Science: The Pluto Flyby of New Horizons

On July 14, 2015, NASA’s New Horizons spacecraft made its closest approach to Pluto and its system of five satellites. Aboard New Horizons were a suite of instruments designed to study Pluto’s atmosphere in greater detail than could be made from Earth. The instruments used to study Pluto’s atmosphere included REX, the Radio science EXperiment (Tyler et al., 2008); Alice, the ultraviolet (UV) spectrograph (Stern et al., 2008); LORRI, the LOng Range Reconnaissance Imager (Cheng et al., 2008); and Ralph (Reuter et al., 2008), a color imager (Multispectral Visible Imaging Camera; MVIC) and near-infrared spectrograph (Linear Etalon Imaging Spectral Array; LEISA). The planned observations for these instruments included the radio occultation using REX, the UV occultation using Alice, and direct imaging of the haze and a search for clouds using LORRI and the MVIC component of Ralph. While each instrument was designed to carry out observations where its strengths were key, the remaining instruments had capabilities that could provide additional supportive observations (Table 1).

Table 1. New Horizons Entry and Exit Profile Information



Location (REX/radio)

193.5° E, 17.0° S

15.7° E, 15.1° N

Location (Alice/UV)

195.3° E, 15.5° S

13.3° E, 16.5° N


Enriched N2, CO



–1.8 km


Local time

16.52, sunset

4.70, sunrise

Near surface T inversion gradient

Upper atmosphere gradient

38.9±2.1 K

+9.5±1.2 K km1

–0.14±0.05 K km1

51.6±3.8 K

+3.9±1.0 K km1

–0.13±0.06 K km1

Cold boundary

<3.5 km


Throughout the article, a discussion is provided about the instrument, the observations it made, and the findings that have been reported thus far.

Radio Occultation/REX

The lowest few scale heights («/i>100 km altitude) of Pluto’s atmosphere were examined using a radio occultation. The experiment utilized four of NASA’s Deep Space Network (DSN) antennas. Each antenna transmitted a 20 kW signal at 4.2 cm that was timed to pass through Pluto’s atmosphere as New Horizons passed behind Pluto. The REX instrument, which was sensitive to Pluto’s atmosphere down to pressures of ≳ 2 µbar (Tyler et al., 2008), recorded the signal. An in-depth analysis of the radio occultation was the main focus of Hinson et al. (2017).

The radio occultation looked at Pluto’s atmosphere over two distinctly different terrains at different local times. The occultation ingress point, located at 193.5° E, 17.0° S (the coordinate system used by New Horizons uses ecliptic north and increasing longitude), occurred at local sunset. The entry point location was over the southeast boundary of Sputnik Planitia, the ice-filled basin that makes up the western lobe of the heart-shaped feature on Pluto. Compared to the majority of Pluto’s surface, the surface composition of Sputnik Planitia is known to be enriched in N2 and CO (Grundy et al., 2016; Protopapa et al., 2017). In contrast, the occultation exit point, located at 15.7° E, 15.1° N, occurred at local sunrise over the near sub-Charon hemisphere. The surface composition of this region has not yet been determined from New Horizons data but it is likely similar to Pluto’s average surface composition (i.e., less N2 and CO, and more CH4 compared to Sputnik Planitia). In addition, stereo imaging of the surface showed the elevation at the entry point is –1.8 km below the average surface elevation of Pluto (for a radius of 1,186.5±1.6 km). Similar stereo imaging over the exit point was not obtained. The difference in surface pressure seen between the entrance and exit point suggests a 5 km difference in Pluto’s radius (Hinson et al., 2017).

The radio occultation focused on the lower 100 km of Pluto’s atmosphere (pressures ≳ 1 µbar). These were the first observations capable of determining the temperature, pressure, and number density distribution profiles down to 1 km above the surface. As discussed in the section “Uncovering the Mysteries of the Atmospheres of Distant Icy Worlds: Stellar Occultations, Spectra, and Early Models of Pluto,” stellar occultations observed from Earth at visible or near-infrared wavelengths only get to a Pluto radius of ≳1,215 km, which now is known to be <i»25 km above the surface. As Hinson et al. (2017) showed (see Figure 4), there are more differences between the entry and exit profiles as they approach the surface starting at an altitude of 25 km. Hinson et al. (2017) found that the entry profile has a strong temperature inversion layer between 3.5 and 20 km altitude with a gradient of +9.5±1.2 K km1, and a cold boundary layer below 3.5 km with a gradient of –0.5±0.7 K km1 for an average temperature of 38.9±2.1 K. Hinson et al. (2017) note that this surface temperature is close to the N2 saturation temperature. The exit location had a weaker inversion layer that had a gradient of +3.9±1.0 K km1, which extended to the ground where the air temperature was estimated at 51.6±3.8 K (measured at 57.0±3.7 K at 1 km above the surface, 18 K warmer than the entry surface temperature).

Figure 4. Results from the REX occultation of Pluto’s atmospere. Temperature (top row), pressure (middle row), and number density (bottom row) for entrance (left column) and exit (right column) from occultation.

Source: J. Cook.

According to PGCMs, a cold boundary layer will arise over regions like Sputnik Planitia due to the diurnal cycle of N2 sublimation and condensation (Forget et al., 2017). In addition, GCMs predict the cold boundary layer reaches its maximum depth right before sunset (predicted 1.5 km, observed 3.5 km), such as the case over the entry point. Over the 3.2 Earth-day-long Pluto night, condensation will cause the cold boundary layer to vanish and the inversion layer to extend to the surface.

The temperature profile of Pluto’s atmosphere at altitudes ≳25 km (≳1,215 km radius) determined by the REX observations is in good agreement with temperature profiles derived from Earth-based stellar occultation data. REX measured a temperature of 107.1±6.4 and 106.1±6.2 K at 25 km altitude for the entry (193.5° E, 17.0° S) and exit (15.7° E, 15.1° N) points, respectively. The temperature profile constructed by Sicardy et al. (2016) from an occultation that occurred about 2 weeks before the New Horizons flyby of Pluto had a maximum temperature of about 110 K at a similar altitude. Note that the ingress and egress points ranged from 108° E–154° E and 32° N–6° N to 232° E–293° E and 37° S–31° S, respectively. Furthermore, the mean temperature gradient at altitudes ≳ 25 km was –0.14±0.05 K km1 and –0.13±0.06 K km1 at entry and exit, respectively, compared to –0.17 K km1 as seen by Sicardy et al. (2016).

UV Occultation/Alice

Examination of Pluto’s atmosphere at altitudes ≳100 km was performed by stellar occultation using Alice aboard New Horizons and the Sun. The Alice instrument is a compact UV spectrograph sensitive to light from 520 to 1870 Å, with a spectral point spread function of 3–6 Å and a spatial field of view of 6° long. This Alice is one of several nearly identical instruments used on other NASA and ESA missions, such as Lunar Reconnaissance Orbiter ’s LAMP (Lyman Alpha Mapping Project), Rosetta’s Alice, Juno’s UVS, and the planned JUICE (JUputer ICy moons Explorer ) and Europa Clipper ’s UVS instruments. At Pluto, the main objective of the Alice instrument was to observe an occultation of the Sun as its light passed through Pluto’s atmosphere and New Horizons went behind the dwarf planet. The results of this experiment were the focus of Young et al. (2018) and build from the initial results reported in Gladstone et al. (2016). The results of the Alice observations are summarized here.

The Alice occultation entered the atmosphere above 195.3° E, 15.5° S, a location that is on the perimeter of Sputnik Planitia and just northeast of the REX entrance location. The occultation exit point happened above 13.3° E, 16.5° N, just northwest of the REX exit location. Similar to the REX observations, the Alice entry and exit occurred during sunset and sunrise, respectively.

Using the Alice observations, Young et al. (2018) derived the mixing ratio as a function of altitude for CH4 from 80 to 1200 km altitude and extrapolated to obtain the surface mixing ratio. They found that the surface CH4 mixing ratio was 0.28 to 0.35%. This value was lower than the surface mixing ratio measured by Gladstone et al. (2016) (0.60–0.84%). Young et al. (2018) explained that the discrepancy was likely due to the fact that Gladstone et al. (2016) lacked the full data set explored by Young et al. (2018), specifically the CH4 line-of-sight measurements below 200 km. Using the data available, Gladstone et al. (2016) also found a higher eddy diffusion coefficient (Kzz = 7.5 × 105 cm2 s1 at the surface and it asymptotically approached 3 × 106 cm2 s1 at 210 km altitude) and homopause altitude (390 km). In comparison, Young et al. (2018) found an eddy diffusion coefficient of Kzz < 104 cm2 s1 and a homopause altitude «/i>12 km. Chemical models of Pluto’s atmosphere by Wong et al. (2017) derived a similar eddy diffusion coefficient to that of Young et al. (2018). In addition, ground-based observations of Pluto’s atmosphere by Lellouch et al. (2015) yielded CH4 mixing ratios (0.49±0.06 in 2008, 0.32±0.05% in 2012) closer in agreement to Young et al. (2018). However, these observed mixing ratios are much higher than would be predicted from Raoult’s law for CH4 gas over CH4 mixed in N2. instead, the source of CH4 gas is likely the CH4-rich ices (Grundy et al., 2016; Protopapa et al., 2017).

Young et al. (2018) report on the line-of-sight abundances of CH4, C2H2, C2H4, and C2H6 from 80 to 1,200 km, 0 to 600 km, 0 to 650 km, and 40 to 550 km, respectively. They found that the CH4 mixing ratio increased with altitude, reaching 1% at about 400 km, and 10% at 800 km, due to diffusive equilibrium. Young et al. (2018) also found C2H4, C2H2, and C2H6 have maxima or an inflection point in density at 410, 320, and 260 km altitude, respectively, and minima in C2H4 and C2H2 at 200 and 170 km altitude, respectively. C2H6, which was the most difficult to measure, also appeared to have a minimum in the 170 to 200 km range. Many of the findings from the UV occultation were also in agreement with the chemical models of Wong et al. (2017) due to condensation of these species.

The work of Young et al. (2018) confirmed the low temperatures in Pluto’s upper atmosphere first suggested in Gladstone et al. (2016). By using the N2 line-of-sight abundance from Alice (900–1,100 km for N2) and REX (0–100 km), Young et al. (2018) derived a temperature profile that ties the two observations together while avoiding discontinuities in temperature, pressure, number density, and their derivatives. Their temperature profile is shown in Figure 5. It should be noted that between the top of the REX observations (100 km) and the bottom of the Alice observations (900 km) for N2, there is a minimum in the temperature profile at about 470 km. According to Young et al. (2018), this minimum was necessary to tie the two observations together and argued it was consistent with the temperature profile inferred by Lellouch et al. (2017) based on ALMA observations of CO and HCN. Chemical models show CO, C2H2, and HCN do not sufficiently cool the atmosphere (Forget et al., 2017; Gladstone et al., 2016; Strobel & Zhu, 2017), and H2O has been a suggested candidate (Strobel & Zhu, 2017). While New Horizons detected H2O ice on Pluto’s surface (Grundy et al., 2016; Protopapa et al., 2017), it will not sublimate at Pluto’s surface temperature. Instead, H2O could be introduced to Pluto’s atmosphere from incoming Kuiper belt material. A second possibility is haze, suggested by Zhang et al. (2017), who argue that haze particles have substantially larger solar heating and thermal cooling rates than gas molecules, dominating the atmospheric radiative balance from the ground to an altitude of 700 km, above which heat conduction maintains an isothermal atmosphere.

Figure 5. Temperature profile derived by Young et al. (2018) shown at four different scales. The temperature profile in the lower atmosphere is shown with the results from REX (REX entrance, dots, and gray shaded region; REX exit, stars bounded by dotted lines; Hinson et al., 2017). Altitude assumes a reference radius of 1,190 km. Major process and atmospheric layers are indicated. Vertical lines on the right indicate the altitude range for which each species was measured. The temperatures in the atmosphere are high because of CH4 heating. The surface is cold because of the ices, radiating in the IR, which results in the gradient in the first tens of kilometers.

Source: J. Cook

Haze and Cloud Observations/LORRI and Ralph

New Horizons’ encounter with Pluto made the presence of haze in its atmosphere irrefutable (Cheng et al., 2017; Gladstone et al., 2016; Stern et al., 2015). Direct imaging at high phase angles from LORRI (Cheng et al., 2008) and the MVIC (Multi-spectral Visible Imaging Camera) component of Ralph (Reuter et al., 2008) revealed the atmosphere has multiple haze layers Gladstone et al., 2016). LORRI is a narrow-angle, high-resolution, 20.8-cm Ritchey-Chrétien telescope. LORRI’s 1,024×1,024 pixel detector has a bandpass that covers the 350 to 850 nm range. Ralph’s MVIC component has four color filters: blue (400–550 nm), red (540–700 nm), near infrared (780–975 nm), a narrow CH4 band (860–910 nm), and two panchromatic filters (400–975 nm). The goal of the high-phase LORRI and MVIC observations of Pluto’s atmosphere was to search for hazes and clouds at vertical resolutions of < 5 km. Discussion about the haze particles based on New Horizons’ observations was the main focus of Cheng et al. (2017) and Gao et al. (2017), while the discussion about possible cloud detection can be found in Stern et al. (2017). In this section, their findings are summarized.

Cheng et al. (2017) reported the detection of haze in LORRI images with phase angles 20° to 169° and spatial resolutions between 90 m/pixel and 4 km/pixel. Prior to the encounter, Pluto’s and Triton’s atmosphere were assumed to be similar.

Photochemical models predicted that a haze layer would be confined to «/i>15 km altitude because the atmosphere above was thought to be too warm to allow the condensation of hydrocarbons or nitriles. However, as shown by Hinson et al. (2017) and Young et al. (2018), the atmosphere of Pluto is much cooler at high altitudes (see Figure 5). As a consequence of the cooler atmosphere, hazes can form at much higher altitudes and lower atmospheric pressures, and these conditions are more akin to the conditions experienced by Titan’s detached haze layer (Cheng et al., 2017).

New Horizons’ images reveal that there are approximately 20 distinct, although sometimes discontinuous haze layers detected in LORRI and MVIC images that extend 200 km above Pluto’s surface. Extinction, attributed to haze, was also detected in the UV occultation observation up to 350 km, but the lack of UV cross-section measurements for the haze particles made determining the abundance complicated (Young et al., 2018). Starting from an altitude of 3 km above the surface, haze layers were seen to have a horizontal extent that spanned several hundred kilometers, with thicknesses of a few kilometers and 10 km spacing (Cheng et al., 2017). Vertical disturbances (e.g., discontinuous haze layers) in the haze layers appeared to be related to surface topography, as was seen in the haze layers that crossed over Cthulhu Macula. Also seen in Pluto’s atmosphere were two zones that had significantly less haze. A lower zone at 30 km altitude was only seen in the more northern latitudes from longitudes 250 to 360°, while an upper zone at 75 km altitude appeared to be global (Cheng et al., 2017). The position, spacing, and thickness of the haze layers could be explained by orographic gravity waves (Cheng et al., 2017). The model also shows the clear zone near 75 km altitude corresponds to a minimum in the Brunt-Väisälä frequency (the frequency at which a vertically displaced parcel of gas will oscillate within a statically stable environment).

The haze appeared blue at visible wavelengths with strong forward scattering properties. While the former physical property suggests Rayleigh scattering of small particles (r ∼ 0.01 µm) to produce the color, the later physical property suggests a larger particle (r > 0.1 µm), possibly an aggregate of smaller spherical particles (Gladstone et al., 2016). Both Cheng et al. (2017) and Gao et al. (2017) concluded that aggregate particles best matched the observations. However, if spherical particles are assumed, then the mass of particles needed to match the observations was greater than the mass production rate of haze particles from CH4 photolysis. Each group found that spherical monomer particles with 10 nm radius and a fractal dimension of 2 yields the aggregate particles where a majority have a bulk radius within the range of 0.08–0.30 µm at 45 km. Cheng et al. (2017) noted that the weak backscattering properties of aggregate particles agreed with observations at altitudes <i»45 km, but not at lower altitudes, thereby suggesting the particle size or shape varied with altitude.

Cheng et al. (2017) suggested spherical particles of 0.5 µm exist at lower altitudes, while Gao et al. (2017) suggested the bulk radius increased with decreasing altitude, but still did not exceed 1 µm. Particle size may increase due to condensation of HCN, C2H2, C2H4, and C2H6, all of which are known to be present in Pluto’s atmosphere (Gladstone et al., 2016; Lellouch et al., 2017; Young et al., 2018). Gao et al. (2017) showed that these species condensation rate peaks in Pluto’s atmosphere at 200 to 400 km altitude and may account for the decrease in density of the latter three species at this altitude, seen in UV occultation (Young et al., 2018).

Cheng et al. (2017) examined the haze images in order to better understand and search for changes over time and distribution. By extracting and comparing brightness profiles of Pluto’s haze layer over the same geographic location over time spans of up to 5.5 hours, Cheng et al. (2017) searched for temporal variations. They concluded that changes seen in the brightness profiles at low altitude were consistent with changes in phase angle due to strongly forward scattering particles. In contrast, the temporal variations seen at higher altitudes (≳ 100 km) were real, but no evidence was found that the haze layer altitude changed over the time. Cheng et al. (2017) found the global distribution of the haze was asymmetric, with greater extinction (brighter in high phase images) seen in the northern hemisphere than at near the equator or southern hemisphere by a factor of 2 to 3. In addition, comparison of brightness profiles obtained at the same time, but different latitudes, showed the haze layers appeared more pronounced at equatorial latitudes.

Because the vapor pressure for several species (HCN, C2H2, C2H6, CH4, and C2H4) in Pluto’s atmosphere can reach saturation levels at altitudes S15 km, a search for cloud features was performed by Stern et al. (2017). Using both MVIC and LORRI images, Stern et al. (2017) found seven possible cloud candidates. All cloud candidates were chosen because they appeared bright relative to the local terrain, were diffuse in appearance, and had a discrete areal extent (10s of km in maximum dimension). Cloud candidates were only seen in high phase angle images (<i»157°) and near the dawn and dusk terminator. The altitude of the cloud candidates cannot be well constrained because stereo imaging for the candidates does not exist, there is no detection of surface-lying shadows, nor does limb viewing geometry help to demonstrate that the features are above the surface. Stern et al. (2017) note that the formation of clouds requires both nucleation sites, which can come from the haze, and condensable material. In order for the material to condense on reasonable timescales (< 106 years to grow a 2 µm radius particle near the surface), condensation must occur at higher altitudes. Stern et al. (2017) suggest condensation happens around 8-1 km altitude where the temperature is ≳60 K and timescales for condensation are short compared to a Pluto year. Stern et al. (2017) conclude that Pluto’s atmosphere is almost entirely free of clouds.

Additionally, LORRI imagery was used by Telfer et al. (2018) to identify dunes on Pluto’s surface. From the dune characteristics, these authors calculated near-surface wind speeds of 1–10 m s1.


New Horizons monumentally advanced the understanding of Pluto’s atmosphere via a combination of radio occultation, UV occultation, and direct imaging. However, those observations only provided a snapshot in time and location and it will take a dedicated Pluto orbiter or even a Pluto lander to truly understand some of the new questions that arose from New Horizons’ observations. The UV occultation demonstrated that CH4-rich ice contribute more to the CH4 mixing ratio than CH4-poor ice (Young et al., 2018). With the exception of HCN, no other nitriles have been detected in Pluto’s atmosphere, although chemical modeling would predict there to be other such as H3CN (Wong et al., 2017). Finally, what exactly is the composition of Pluto’s haze? New Horizons lacked the spectral resolution to determine the haze composition.

Continued stellar occultation and spectral measurements are necessary to track how Pluto’s atmosphere is changing with time (especially in terms of total surface pressure and relative abundances of constituents). Progress must also continue to be made to build coupled surface-atmosphere models that simulate Pluto’s atmosphere in 3-D and time over multiple Pluto years, which will hopefully become an easier task as computer processing speed inevitably becomes faster. This type of model will shed further light on Pluto’s wind velocities (which cannot be observed remotely) and the distribution of ice on Pluto’s surface.

Similarly for Triton, models should be used to address the following questions: What is the atmospheric role in surface character (i.e., surface streaks and other patterns), and vice versa (i.e., surface plumes creating clouds and sublimation and deposition driving atmospheric flow)? In what ways do large-scale atmospheric circulations impact the otherwise axisymmetric flux of volatile materials to surface deposition regions? In what ways does the atmosphere control the temperature, nature, and physical properties of N2 ice across Triton? In what ways do the spatial distributions of albedo, emissivity, and other surface characteristics affect the atmosphere? What is the thermal structure and surface pressure distribution of the lower atmosphere, and what role does atmospheric dynamics play in maintaining them?

What are the speed, spatial extent, and temporal evolution of the circulation directly driven by deposition and sublimation? What is the wind and temperature structure that causes the observed behavior of the sheared-off plumes and clouds and the general directionality of surface albedo streaks?

Furthermore, there is an urgent need for more observations of Triton, as conditions permit. The Outer Planet Assessment Group (OPAG) has stated, “the Ice Giants, Uranus and Neptune, remain high-priority targets for exploration as reflected in the Visions and Voyages Planetary Science Decadal Survey” (Findings, 2017). While New Horizons was able to observe Pluto’s full disk (albeit at varying resolution), Voyager 2 did not observe Triton’s full disk. The observations are also separated in time by 26 years, and newer instrument technologies would improve the quality of flyby observations of Triton. Now that it has been shown that Pluto and Triton show significant differences (e.g., plumes on Triton but not Pluto, well-defined haze layer on Pluto), comparative studies of these bodies is important to determine how bodies in similar locations evolved to have different characteristics.

Further Reading