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: 04 December 2020

Tectonism of Mercuryfree

  • Paul K. ByrnePaul K. ByrneNorth Carolina State University, Department of Marine, Earth, and Atmospheric Sciences


Mercury, like its inner Solar System planetary neighbors Venus, Mars, and the Moon, shows no evidence of having ever undergone plate tectonics. Nonetheless, the innermost planet boasts a long record of tectonic deformation. The most prominent manifestation of this history is a population of large scarps that occurs throughout the planet’s cratered terrains; some of these scarps rise kilometers above the surrounding landscape. Mercury’s smooth plains, the majority of which are volcanic and occupy over a quarter of the planet, abound with low-relief ridges. The scarps and ridges are underlain by thrust faults and point to a tectonic history dominated by crustal shortening. At least some of the shortening strain recorded by the ridges may reflect subsidence of the lavas in which they formed, but the widespread distribution of scarps attests to a planetwide process of global contraction, wherein Mercury experienced a reduction in volume as its interior cooled through time.

The onset of this phenomenon placed the lithosphere into a net state of horizontal compression, and accounts for why Mercury hosts only a few instances of extensional structures. These landforms, shallow troughs that form complex networks, occur almost wholly in volcanically flooded impact craters and basins and developed as those lavas cooled and thermally contracted. Tellingly, widespread volcanism on Mercury ended at around the same time the population of scarps began to form. Explosive volcanism endured beyond this point, but almost exclusively at sites of lithospheric weakness, where large faults penetrate deep into the interior. These observations are consistent with decades-old predictions that global contraction would shut off major volcanic activity, and illustrate how closely Mercury’s tectonic and volcanic histories are intertwined.

The tectonic character of Mercury is thus one of sustained crustal shortening with only localized extension, which started almost four billion years ago and extends into the geologically recent past. This character somewhat resembles that of the Moon, but differs substantially from those of Earth, Venus, or Mars. Mercury may represent how small rocky planets tectonically evolve and could provide a basis for understanding the geological properties of similarly small worlds in orbit around other stars.


Its status as the innermost planet means that Mercury has been the least-studied rocky body in the inner Solar System. Situated too close to the Sun to be easily observed with telescopes from Earth, this enigmatic world is sufficiently deep inside the star’s gravity well that any visiting spacecraft must spend years shedding speed before successfully making orbit—a constraint not applicable to missions to the Moon, Mars, or even Venus. However, a full understanding of how the planets in this and other star systems came to be requires that Mercury be explored as much as those other worlds. By investigating the planet’s surface, interior, exosphere, and magnetosphere, the properties and processes unique to Mercury and its evolutionary path, as well as those common to silicate planetary bodies in general, can be better identified. Moreover, with its anomalously large core mass fraction (Johnson & Hauck, 2016), Mercury may provide a basis for understanding extrasolar planets of similar size (Barclay et al., 2013) or with comparable interior structure (Santerne et al., 2018).

A key expression of a planet’s geological evolution is the inventory of tectonic landforms recorded on its surface. The tectonic characteristics of a planet reflect the interplay of volcanism, chemistry, interior structure, and thermal evolution. The planet’s geological history can thus be more fully comprehended by examining the nature and formational histories of tectonic structures and their spatial and temporal relationship to the interior. Comparing the structures present on Mercury with those on other worlds, including Earth, helps efforts to determine how rocky planets behave in general.

This article summarizes the current understanding of the tectonic character of Mercury. First, the state of knowledge following the first spacecraft to visit Mercury, NASA’s Mariner 10, is presented; the types and distributions of landforms observed across the ~45% of the planet surface imaged by this mission are reviewed, and prospective formation mechanisms discussed. This view prevailed from the mid-1970s until the first flyby of the NASA MESSENGER mission in 2008, which went on to enter orbit around Mercury in 2011 and operated there until 2015. MESSENGER dramatically reshaped our understanding of the planet, and the major findings and implications from this mission constitute the bulk of the article. Finally, several important outstanding questions are considered in the context of the subsequent phase of Mercury exploration by the joint ESA–JAXA BepiColombo mission. The article concludes with a synthesis of Mercury tectonism as it stands, and what insight this diminutive planet can provide of the geological and thermal evolution of rocky bodies in this and other solar systems.

The View Before MESSENGER

The Mariner 10 mission launched from Earth in November 1973 and made a total of three flybys of Mercury in 1974 and 1975. Because of the geometry and period of its heliocentric orbit, Mariner 10 saw the same hemisphere of Mercury in each of its flybys, ultimately returning images of a little less than half of the surface. Yet even without performing a global survey of the planet, Mariner 10 led to a dramatic new understanding of Mercury.

Mariner 10 Observations

Images returned by the Mariner 10 Television Photography Experiment (Murray et al., 1974) showed a planetary surface similar to that of the Moon, battered by impact craters and without the major volcanic constructs so prevalent on Earth, Venus, and Mars. Variations in surface texture and areas with far fewer craters than others implied resurfacing to some degree, possibly by volcanic activity (Strom, Trask, & Guest, 1975). Mariner 10 observations provided evidence that widespread tectonic deformation had occurred across the imaged hemisphere (Murray et al., 1974; Strom et al., 1975).

This deformation was most prominent within an expansive region of relatively smooth terrain termed Caloris Planitia (Figure 1). These plains, situated within the largest preserved impact basin on Mercury, were seen by Mariner 10 to host complex networks of intersecting, linear troughs and ridges. The system of troughs, with some examples tens of kilometers long, were interpreted as extensional tectonic structures because of their morphological resemblance to graben (down-dropped crustal blocks bounded on each side by inward-dipping normal faults) observed on the Moon and Mars (Strom et al., 1975). Similarly, the ridges were regarded as evidence for crustal shortening on the basis of comparison with landforms termed wrinkle ridges common to the vast lunar and Martian lava deposits (Strom et al., 1975). Wrinkle ridges likely reflect some combination of folding and thrust faulting (e.g., Mueller & Golombek, 2004), albeit to a modest extent. Together, these populations of extensional and shortening structures were taken to indicate dominantly vertical tectonics within the basin that hosts Caloris Planitia, with uplift from isostatic adjustment the mechanism responsible for the graben and subsidence from volcanic loading by the plains the means by which the ridges formed (Dzurisin, 1978; Strom et al., 1975).

Figure 1. Caloris Planitia as seen by Mariner 10 (lefthand side of the image). These plains, situated in an impact basin more than 1,500 km in diameter (Murchie et al., 2008; the rim of which is marked by white arrows), were observed to have relatively fewer craters per unit area than much of the rest of the planet surface. Note the substantial number of ridges and grooves within Caloris Planitia; no other region on Mercury is as tectonically complex. The scale given in this figure is approximate; illumination is from the right. The image is unprojected and is from computer mosaic PIA03102 (NASA/JPL).

Additional instances of crustal shortening were identified outside of Caloris Planitia, but with dimensions substantially greater than those of the wrinkle ridges. These larger structures were called “lobate scarps” by the Mariner 10 team on the basis of their arcuate outline, and were observed throughout the heavily cratered terrains on Mercury (Strom et al., 1975) (Figure 2). With rounded crests, the possibility arose that these landforms were thick lava flow fronts. But observations of scarps cross-cutting a variety of surface units, and offsetting crater rims in a manner consistent with thrust faulting, strongly suggested that theirs was a tectonic origin. This interpretation was supported by the great sizes of many individual scarps, which in some cases were hundreds of kilometers long and more than 1 km in height (Strom et al., 1975). A morphologically similar set of prominent, linear landforms, termed high-relief ridges, was identified, and accorded a similar formation mechanism. Given the distribution of these large landforms throughout the entire hemisphere seen by Mariner 10 (Trask & Guest, 1975), it became apparent that they must represent a mechanism that operated globally (Strom et al., 1975). Although numerous processes can lead to crustal shortening, the best explanation for a planetwide population of thrust faults is from global contraction (Strom et al., 1975; Solomon, 1978).

Figure 2. An example of a lobate scarp as seen by Mariner 10. Two craters (A and B) are cut by a scarp over 130 km long that trends northwest, and appear to have been shortened in a southwest direction. Observations such as these gave rise to the view that lobate scarps, rather than being volcanic in nature, were tectonic structures reflecting substantial crustal shortening. Illumination in this unprojected image is from the left. After Strom et al. (1975).

This phenomenon arises from the cooling through time of a planetary body’s interior, with that cooling resulting in a decrease in volume and proportionate reduction in surface area. This surface areal reduction is, in turn, accommodated by thrust faults that could plausibly account for the scarps seen so widely across Mercury. Estimates by the Mariner 10 team of the change of volume of Mercury, given as a function of decrease in radius, spanned a range of about 1–2 km (Strom et al., 1975), although with considerable uncertainty in these values. Notably, such values were substantially below those predicted by thermal evolution models for Mercury, which generally called for radius reduction values of ~5–10 km (e.g., Hauck, Dombard, Phillips, & Solomon, 2004; Schubert, Ross, Stevenson, & Spohn, 1988; Solomon, 1977). In any case, assessments of age relations between lobate scarps and the surface units they deform pointed to the majority of crustal shortening taking place near the end of the Late Heavy Bombardment (Strom et al., 1975).

The results of the Mariner 10 mission indicated that Mercury tectonically resembles Mars and the Moon, with a lithosphere consisting of a single shell and no evidence of the mosaic of mobile tectonic plates that characterizes Earth (i.e., Mercury is a “one-plate planet,” Solomon, 1978). But this initial appraisal did not spell the end of interest in Mercury’s tectonic character. Key questions remained unanswered: What did the hemisphere not imaged by Mariner 10 look like? Were the lobate scarps documented by Mariner 10 truly globally distributed? Did they reflect the global contraction of Mercury and, if so, to what extent had the planet’s volume been reduced? The latter question in particular was especially important because an accurate measure of the reduction in planetary radius is tied to interior structure, the bulk abundance of heat-producing elements in the crust and mantle, the prospect for mantle convection through time, and even the history of cooling and present-day structure of Mercury’s large iron core, the source of its intrinsic magnetic field (e.g., Grott, Breuer, & Laneuville, 2011; Hauck et al., 2004; Michel et al., 2013; Strom et al., 1975; Tosi et al., 2015). And what of the planet’s extensional structures? Aside from that within Caloris Planitia, virtually no extensional tectonic deformation was observed by Mariner 10, a surprising finding given the widespread occurrence of troughs and graben on the Moon, Mars, and Venus.

These and other questions persisted long after the Mariner 10 mission ended. Analysis of the tectonic characteristics of Mercury continued, leading to revised estimates of the planet’s change in radius (e.g., Watters, Robinson, & Cook, 1998), although these estimates remained consistently below those predicted by thermal history models for Mercury (e.g., Dombard & Hauck, 2008; Hauck et al., 2004). The limited resolution and spatial coverage of Mariner 10 image data challenged efforts to fully understand the geological, thermal, and tectonic history of the planet, a problem that persisted until the first visit by the MESSENGER spacecraft in January 2008.

The View After MESSENGER

The NASA MErcury Surface, Space ENvironment, GEochemistry, and Ranging (MESSENGER) mission (Solomon et al., 2008) launched in August 2004 and, after a cruise phase that lasted six and a half years and involved flybys of Earth, Venus, and Mercury itself, arrived at its destination in March 2011. The spacecraft was equipped with an instrument payload far more capable than that of Mariner 10; in addition to an imaging system that included narrow- and wide-angle cameras (NAC and WAC, respectively), MESSENGER carried a laser altimeter and several spectrometers, and was also capable of performing magnetic and radio science measurements (Cavanaugh et al., 2007; Goldsten et al., 2007; Hawkins et al., 2007; Schlemm et al., 2007; Solomon et al., 2008). The Mercury Dual Imaging System (MDIS) and the Mercury Laser Altimeter (MLA) instruments together allowed for the acquisition of topographic data.

The trajectory that MESSENGER followed to ultimately orbit Mercury required three flybys of the innermost planet, and the spacecraft was able to acquire data during each close approach (Solomon et al., 2008). By the third flyby almost the entire planet surface had been imaged, but it was the orbital phase of the mission (lasting a little more than four Earth years) that enabled the major advances in understanding of the planet’s geology, its interior, composition, and exosphere and magnetosphere. Ultimately, the observations taken by MESSENGER led to the creation of several global base maps with a variety of solar incidence and illumination azimuths (Chabot et al., 2016), thousands of high-resolution MDIS NAC images, a global digital elevation model (DEM) generated from the control network derived from the development of the global image base maps (Becker et al., 2016), and higher-resolution DEMs created for select regions with stereophotogrammetric techniques (e.g., Fassett & Crowley, 2016). Individual laser altimetric profiles from MLA were also combined to form an interpolated DEM of the northern hemisphere (e.g., Zuber et al., 2012). These data sets underpin studies of the planet’s tectonic properties (e.g., Rothery & Massironi, 2010; Di Achille et al., 2012; Byrne et al., 2014) in the MESSENGER era and beyond.

Mercury’s Tectonic Inventory

A key advance offered by MESSENGER was the ability to characterize the global inventories of shortening and extensional tectonic structures across Mercury (Figure 3). As had been inferred from Mariner 10 observations, shortening structures are distributed across the entire planet surface, deforming all major surface units (e.g., Byrne et al., 2014). Extensional landforms morphologically similar to those within Caloris Planitia (the volcanic origin of which was confirmed) were observed elsewhere on Mercury, although overwhelmingly situated within solidified volcanic flows. And a curious set of long-wavelength warps, not previously recognized on the planet was also seen by MESSENGER. The origin of these warps remains unknown, but they may have a tectonic origin and so are included in the following discussion.

Figure 3. The shortening and extensional tectonic landforms of Mercury. Shortening structures are shown in warm colors, whereas extensional landforms are shown in cool colors. Long-wavelength topographic undulations are shown with black lines; arrows denote downslope directions. The global population of shortening structures is from Byrne et al. (2014); extensional landforms are from Klimczak et al. (2012) and Ferrari et al. (2015). The map is in a Robinson projection centered at 0°E; the graticule is in 30° increments in latitude and longitude. After Byrne et al. (2018).

The terms wrinkle ridge, lobate scarp, and high-relief ridge arose from Mariner 10 observations (Strom et al., 1975; Melosh & McKinnon, 1988; Watters, Robinson, Bina, & Spudis, 2004; Watters & Nimmo, 2010). Wrinkle ridge morphology is characterized by a broad, steep-sided but low-relief arch that is symmetric in cross-section (Figure 4A), variously with or without a crenulated crest; in contrast, the larger lobate scarps have a resolvable asymmetric form in cross-section (Figure 4B). High-relief ridges share the cross-sectional symmetry of wrinkle ridges with the scale of lobate scarps. To first order, MESSENGER data confirmed earlier findings that wrinkle ridges are situated within smooth plains units, whereas lobate scarps and the comparatively scarcer high-relief ridges deform the older, more cratered terrain (Byrne et al., 2014; Byrne, Klimczak, & Şengör, 2018; Watters & Nimmo, 2010).

Figure 4. Examples of shortening landforms on Mercury. (A) Smooth plains structures in Mercury’s northern plains; note the localization of shortening landforms at the rims of buried craters (shown by white arrows). (B) Carnegie Rupes, an example of a relatively linear cratered plains structure that is almost 270 km long and that cuts through the 133-km-diameter Duccio crater. These images are taken from the global morphology base map (Chabot et al., 2016) and are in azimuthal equidistant projections, centered as follows: (A) 60.3°N, 52.9°E; (B) 58.5°N, 306.7°E. After Byrne et al. (2018).

A consensus firmly exists that these structures all reflect crustal shortening via some combination of thrust faulting and folding. In many instances, however, attempts to catalog shortening structures on the basis of morphology falls short. This is because these structures commonly change shape, size, and inferred fault dip direction along their length, such that wrinkle ridges frequently take on the asymmetric form of lobate scarps as they deform smooth plains, some lobate scarps boast archlike shapes and so resemble enormous wrinkle ridges, and many structures adopt one form along their length before switching to another. Moreover, no criterion yet exists by which high-relief ridges can be uniquely and systematically distinguished from lobate scarps, especially as one form can transition into the other (Byrne et al., 2018). To that end, therefore, the most comprehensive survey to date of shortening structures on Mercury described landforms on the basis of the geological setting in which they occur (Byrne et al., 2014).

Under this scheme, the most common type of shortening landform is the smooth-plains structure, which deform the eponymous plains units (Denevi et al., 2013) that occupy more than a quarter of the planet surface (Figure 3). These structures, the majority of which would otherwise be classified as wrinkle ridges but include many landforms with a lobate scarp-like morphology, represent 63% of the total number of mapped shortening features but possess only about 50% of the cumulative mapped length (Byrne et al., 2014). A plurality of smooth-plains structures is situated in the largest expanse of volcanic smooth plains on Mercury, the vast Borealis Planitia (cf. Head et al., 2011). The remainder of this type of structure occurs within Caloris Planitia (Murchie et al., 2008; Denevi et al., 2013) and inside other volcanically infilled large impact basins and craters. Cratered-plains structures are distributed across most of the rest of the planet (Figure 3). Constituting around 31% of the total number of mapped shortening structures but 40% of total mapped length, these landforms are generally larger than the smooth-plains structures (Byrne et al., 2014).

A subset of both smooth- and cratered-plains structures constitute a third group of shortening landform: those that either demarcate partially to entirely buried craters in volcanic smooth plains (i.e., “ghost craters”) or that deform the volcanic infill of large basins (e.g., Fegan et al., 2017). Termed basin- and crater-related structures by Byrne et al. (2014; Figure 3), these structures represent 4% of the total number and 5% of the total length of all mapped shortening landforms on the planet. A fourth grouping of shortening features by Byrne et al. (2014) comprises high-terrain-bounding structures, which represent 2% of the total number but a disproportionately high 6% of mapped length of such features. High-terrain-bounding structures lie along the margins of, and are oriented such that the component thrust faults dip under, locally elevated parts of the planet surface. These structures are among the longest and tallest recognized on Mercury; for example, the 820-km long Enterprise Rupes thrust fault system, which crosses the 720 km -diameter Rembrandt basin, stands almost 3 km above the basin interior in parts.

Instances of crustal extension on Mercury are more widespread than observations from Mariner 10 suggested but, notably, are restricted almost exclusively to sites of ponded lavas and impact melt (Figure 3). The elaborate system of troughs and graben within Caloris Planitia seen by the earlier mission deforms the entire interior of this basin and is dominated by a remarkable network of radially oriented graben named Pantheon Fossae (Murchie et al., 2008; Watters et al., 2009; Figure 5A). A similar if less complex set of graben and troughs occurs within the volcanically infilled Rembrandt basin (e.g., Ferrari et al., 2015), and intersecting patterns of graben characterize the lava-filled interiors of several midsize basins, including Mozart, Rachmaninoff, and Raditladi (Blair et al., 2013; Prockter et al., 2010). Smaller extensional structures (probably normal faults) are present within ghost craters across Borealis Planitia (Klimczak et al., 2012; Figures 3 and 5B), as well as within ponded impact melt deposits in relatively well-preserved craters such as Degas (although these structures may be a combination of normal faults and joints: Byrne et al., 2018). The only setting where extensional structures have been observed on Mercury outside of smooth plains units is along the crests of large shortening landforms (Banks et al., 2015).

Figure 5. Examples of extensional landforms on Mercury. (A) A portion of Pantheon Fossae, the radial network of graben within the Caloris basin that originates near the center of Apollodorus crater. (B) Graben with multiple orientations inside two ghost craters (the rims of which are marked by white arrows), themselves situated within the 317-km-diameter Goethe basin. Note several additional graben immediately south of the larger of the two ghost craters. These images are taken from the global morphology base map (Chabot et al., 2016) and are in azimuthal equidistant projections, centered as follows: (A) 30.0°N, 161.0°E; (B) 81.0°N, 309.0°E. After Byrne et al. (2018).

Akin to the other one-plate planets, Mercury shows no evidence of the major lateral motion associated with tectonic plate boundaries (Byrne et al., 2018). Minor instances of strike-slip deformation are present, however, though generally as secondary structures associated with orthogonal shortening or extension. For instance, left- and right-lateral ramps were recognized along the margins of the highly arcuate Beagle Rupes shortening structure (Rothery & Massironi, 2010), and similar features were noted along other major thrust systems across the planet (Galluzzi, Di Achille, Ferranti, Popa, & Palumbo, 2015; Massironi et al., 2015).

Perhaps the most unusual component of Mercury’s tectonic inventory is a set of long-wavelength topographic undulations identified in the northern hemisphere. These features were first observed in DEMs derived from MDIS flyby images (Oberst et al., 2010), before being confirmed by MLA altimetry during orbital operations (Zuber et al., 2012). These warps are most prominent as they cross Caloris Planitia in an approximately east–west direction (Figure 6), although they continue into the more cratered terrain to the west of those plains. The undulations are much broader than they are tall, with wavelengths of 850 to 1,120 km, and amplitudes of up to 2.5 to 3 km (Klimczak et al., 2013) and, as a result, are resolvable only with topographic data. Nonetheless, they are striking landforms: they have no apparent counterpart on any other inner Solar System body, and their modification of the long-wavelength topography on Mercury is such that portions of Caloris Planitia have been elevated above the rim of the basin in which they are hosted (Zuber et al., 2012). A similar topographic warp was also found within Borealis Planitia at high northern latitudes, with a diameter of about 1,000 km and an elevation of around 1.5 km (Klimczak et al., 2012; Zuber et al., 2012). Notably, the warps within and proximal to Caloris and Borealis Planitiae feature volcanically infilled impact craters with floors that are tilted in the same direction as the down-slope trend of the long-wavelength topography (Balcerski, Hauck, & Sun, 2012; Zuber et al., 2012). MESSENGER data indicate that, on the basis of admittance values (the ratio of gravity to topography in the wavenumber domain), the topographic rise within Borealis Planitia is either buoyantly supported near the base of the mantle or is elastically supported by the strength of the lithosphere (James, Byrne, Solomon, Zuber, & Phillips, 2014).

Figure 6. An MLA profile (orange) across the central portion of Caloris Planitia. These plains show sinusoidal topographic undulations approximated with a wavelength of 850 km and an amplitude of 2.5 km (burgundy). The profile, which has a vertical exaggeration of ~100:1, shows both the tilted floor of Atget crater, situated near the center of the plains, and where the highest elevations of these ponded lavas exceed those of the northern basin rim. After Byrne et al. (2018).

The Lithosphere of Mercury

As for the other terrestrial worlds, Mercury possess a rigid but deformable outer shell termed the lithosphere (Phillips, Byrne, James, Mazarico, Neumann, & Perry, 2018). The lithosphere, in turn, consists of a relatively cold, upper region where brittle deformation occurs, and a relatively warm, lower region that responds to stress in a ductile manner (e.g., Kohlstedt & Mackwell, 2010). The types, distributions, and geometries of tectonic structures on Mercury’s surface reflect the stresses that formed them as well as the mechanical properties of the lithosphere itself, and so a full characterization of the tectonic inventory of the planet provides insight into its interior, which is not otherwise easily accessible.

In the brittle lithosphere, tectonic deformation is accommodated by localized fracturing processes, which results in two dominant types of structure: joints and shear fractures (i.e., faults). As discussed earlier, faults are by far the most prominent and largest fractures in the Mercury lithosphere, and the most prominent of those structures are thrust faults. Thrusting occurs when one of two horizontal stress components in the lithosphere, σH, is greater than the stress component acting vertically, σV (usually the overburden stress; i.e., σH > σV > 0), with compressive stress taken to be positive. The global presence of thrust faults on Mercury and their temporal relations with other structures and surface units (see section on “Timing of Deformation”), indicates that this stress state prevailed in the lithosphere for most of the planet’s history, consistent with inferences by the Mariner 10 and MESSENGER mission teams (Byrne et al., 2014; Strom et al., 1975). Extensional deformation is the result of a stress state where σV > σH > 0, leading to the formation of joints, normal faults, and graben. That such tectonic structures are so spatially restricted on Mercury indicates that extensional stresses within the lithosphere were present only locally (see section on “Mercury’s Tectonic Inventory”).

Rock behaves in a ductile manner below the brittle portion of the lithosphere. The change in deformation behavior is termed the brittle–ductile transition (BDT), the depth of which is generally taken to be where the brittle and ductile strengths of the lithosphere are equal (Figure 7). In the ductile regime, deformation is dominantly accommodated not by narrow, discrete zones of failure but by more distributed plastic flow mechanisms such as dislocation glide or diffusion creep (e.g., Kohlstedt, Evans, & Mackwell, 1995). Ductile deformation, although resulting from applied stress just as for brittle failure, is largely controlled by rock type, thermal gradient, and strain rate. Uncertainty in these parameters (since they are not readily measurable from orbit) means that the strength profile of Mercury’s lithosphere and the depth of the BDT have yet to be fully determined.

Figure 7. A possible scenario for the evolution of lithospheric strength and thickness for Mercury’s different thermal environments, shown for the equatorial hot (orange) and cold (red) poles as well as the spin poles (grey) for both extensional and contractional tectonic regimes. The brittle regime is represented with Byerlee’s law (after Klimczak, 2015); the lithospheric thickness and ductile regime are modified after the models of Williams et al. (2011). Lithospheric strength envelopes are shown in 1 Gyr increments after Mercury’s formation, for a horizontal strain rate of 10–19 s–1 and for ductile flow laws appropriate to a dry basaltic crust and a dry dunite mantle with their interface shown here for an exemplar depth of 100 km. After Byrne et al. (2018).

Nonetheless, some understanding of the mechanical structure of the lithosphere can be gleaned with remotely sensed observations. For instance, the morphology of a fault-related landform reflects the geometry of the underlying fault (e.g., Klimczak, 2014; Byrne et al., 2015), and so efforts to characterize the depth to which a fault penetrates offers information on the thickness of the seismogenic layer in which it formed, the thermal gradient at the time of formation, and even the depth of the BDT (e.g., Nimmo & Watters, 2004). Several studies based on both Mariner 10 and MESSENGER data have returned estimates for BDT depth of 30 to 40 km (e.g., Nimmo & Watters, 2004; Egea-González et al., 2012)—although because faults can terminate at weak layers within the brittle lithosphere, and since the transition to ductile deformation has likely deepened through time as the planet has cooled, this range is a minimum bound on the present-day depth of the BDT. It is also possible that variations in surface temperature across the planet (Vasavada, Paige, & Wood, 1999) affect the strength and structure of the lithosphere, with differences of up to 15 km in lithospheric thickness between the hot and cold poles along Mercury’s equator and even greater differences at the rotational poles (Williams, Ruiz, Rosenburg, Aharonson, & Phillips, 2011), with commensurate effects upon the depth of the transition from brittle to ductile failure.

Deformation Mechanisms

The profusion of shortening structures across Mercury strongly argues for a mechanism capable of driving crustal shortening that was global in scale (Strom et al., 1975). Although major shortening systems are prevalent on Earth, they primarily represent the convergence of tectonic plates, a scenario that clearly has not operated on Mercury. And whereas localized processes can create shortening structures (e.g., such as basal thrust faults that result from landslides or volcano spreading; see Borgia, Delaney, & Denlinger, 2000), or thrusts arising from lava subsidence (e.g., Zuber & Mouginis-Mark, 1992), few mechanisms are capable of generating a planetwide population of thrust faults.

For this reason, Strom et al. (1975) favored global contraction as the principal means by which Mercury’s population of lobate scarps formed. This view was supported by MESSENGER-era studies (e.g., Byrne et al., 2014; Watters et al., 2016), which have been able to use the global image data sets returned by the orbital mission. Importantly, detailed structural mapping of the planet surface has indicated that, collectively, the lobate scarps and wrinkle ridges represent a decrease in the planetary radius by up to 7 km (Byrne et al., 2014). When considering the scarps alone (i.e., excluding all those shortening structures hosted by Mercury’s smooth plains), that value drops only by about 11% (Byrne et al., 2014). Such a finding is supported by two independent means (Byrne et al., 2014) involving techniques whereby the measured vertical relief and mapped length of select lobate scarps are related to true fault displacement (cf. Clark & Cox, 1996) via assumptions for fault dip angle (typically 25°–35°, although numerous thrust faults on Mercury have lower dip angles [Galluzi et al., 2015], so the values here may underestimate the true extent of the planet’s decrease in radius). Moreover, given the likely mechanical properties of the Mercurian lithosphere, some portion of global contraction-induced strain must have been accommodated elastically; that is, before the formation of actual thrust faults and their associated scarps (Klimczak, 2015). Indeed, depending on the assumed mechanical properties of the lithosphere, Mercury’s radius decreased by 0.4 to 2.1 km before the onset of faulting. It is possible, then, that the planet experienced a radius reduction of as much as about 9 km from interior cooling—far more than estimates derived from Mariner 10 data (e.g., Strom et al., 1975; Watters, Robinson, Bina, & Spudis, 2004), but consistent with decades-old thermal evolution modeling predictions (e.g., Hauck et al., 2004; Dombard & Hauck, 2008; Solomon, 1977).

Much of Mercury’s geological history can be understood through the framework of global contraction, but additional deformation mechanisms have left their mark on the planet. For example, in calculating their set of radius change values, Byrne et al. (2014) excluded Mercury’s population of smooth plains-hosted shortening structures because it is possible that at least some portion of the strain accommodated by these landforms arose from volcanic load-induced flexure or subsidence (e.g., Melosh & McKinnon, 1988). This latter phenomenon in particular has been invoked to explain wrinkle ridges in the lunar maria (e.g., Bryan, 1973), the Olympus Mons caldera (Zuber & Mouginis-Mark, 1992), and those within impact basins across Mars (e.g., Head, Kreslavsky, & Pratt, 2002), and loading of the Martian lithosphere by the Tharsis Rise likely drove wrinkle ridge formation peripheral to that major volcanic center (e.g., Mueller & Golombek, 2004). However, as comparable landforms on Mercury are almost universally situated within late-stage lava units, any subsidence of, or flexure from loading by, volcanic material has been presumably restricted to those units.

Other processes may have contributed to Mercury’s record of crustal shortening, but the extent to which they operated (if at all) is unclear. For instance, a decrease in the rotational rate of Mercury as a result of gravitational interaction with the Sun (i.e., tidal spindown) should have tectonically deformed the surface by the relaxation of a tidal bulge (e.g., Melosh, 1977; Melosh & McKinnon, 1988). Such deformation was proposed to have contributed to the distribution and types of tectonic structure observed across the hemisphere imaged by Mariner 10 (Melosh & McKinnon, 1988), although global tectonic mapping did not find evidence for such a despinning pattern (Byrne et al., 2014). Further, the pattern predicted to result from tidal despinning differs from earlier findings when geophysical models are assessed with rock mechanics (Klimczak, Byrne, & Solomon, 2015). It is possible that despinning played some part in the evolution of Mercury’s tectonic structures, but substantial uncertainty exists as to when this process may have started on Mercury, and thus whether it could have overlapped with the timing of major crustal deformation (Klimczak et al., 2015). Yet other tectonic processes have also been invoked for Mercury, including mantle convection (e.g., King, 2008) and planetary reorientation (e.g., Matsuyama & Nimmo, 2009). However, neither the patterns of strain predicted to result from convection nor reorientation match the mapped distributions of shortening structures across Mercury (Byrne et al., 2014), so any role played in the tectonic history of Mercury by these mechanisms remains an open question.

But what of Mercury’s extensional structures? Because virtually all surviving evidence of such deformation on the planet is restricted to volcanic smooth plains deposits, within or proximal to impact craters and basins, it follows that the processes leading to extensional strains are tied to the properties of volcanic units or impact features. And this appears to be the case: thermal contraction of ponded lava flows within craters and even midsize basins is probably responsible for the distinctive pattern of graben so commonly observed in these settings (Blair et al., 2013; Freed, Solomon, Watters, Phillips, & Zuber, 2009; Klimczak et al., 2012). This mechanism may even apply to the larger, radial sets of graben within Caloris Planitia and the volcanic plains inside the Rembrandt basin, although flexural uplift (Freed et al., 2009) driven by lower crustal flow (Watters, Nimmo, & Robinson, 2005) or the emplacement of major volcanic loads around the basin perimeter (e.g., Kennedy, Freed, & Solomon, 2008), or even dike propagation (e.g., Head et al., 2008), have been proposed to account for the enigmatic Pantheon Fossae. However, the strains represented by these graben correspond to an uplift of as much as 10 km, which is clearly not the case for the Caloris Planitia-hosting basin floor, and these strains also exceed that expected for a radial dike swarm (Klimczak, Schultz, & Nahm, 2010). The origin of Pantheon Fossae, and the similar set of radial graben within Rembrandt, is therefore still unknown.

The origin of the long-wavelength warps on Mercury is no better understood at present. The shapes of some of these landforms (particularly those within and adjacent to Caloris Planitia) matches the pattern of linear rolls predicted by earlier numerical simulations to result from mantle convection (e.g., King, 2008). However, the mantle thickness indicated by MESSENGER (~400 km) is substantially less than the values in those simulations (~600 km), and so convection cells within Mercury would likely not manifest as such a pattern (e.g., Byrne et al., 2018). Folding of the lithosphere was also considered for Mercury, in fact long before MESSENGER results were available but in response to the mismatch between Mariner 10–derived radius changes estimates and those predicted by thermal evolution models (e.g., Dombard, Hauck, Solomon, & Phillips, 2001). Although some model configurations (e.g., those with an elastic–plastic rheology for the lithosphere) allow for folds to develop in response to global contraction (Dombard et al., 2001), more recent simulations with an elastic–viscous–plastic rheological model indicate that such folding is unlikely for Mercury (Kay & Dombard, 2017).

Timing of Deformation

The key to determining when tectonic activity took place on a planetary surface from remotely sensed observations is to search for evidence of superposition and cross-cutting relations. That is, if a tectonic structure deforms a given surface unit, then it follows that the structure must be younger than that unit. This approach has been used widely in planetary science, although unlike landforms such as volcanic flows or impact craters, which form once, tectonic structures can be reactivated by a later source of stress or can even remain active over extended periods of time. Therefore, observations of tectonic cross-cutting relations tell us of the most recent phase of activity.

Nevertheless, there is some understanding of the timing of tectonic processes on Mercury. For example, the planet’s population of thrust faults appears to deform all major surface units (Banks et al., 2015), which places at least the last increment of crustal shortening after the youngest of those surface units was emplaced. Given that shortening structures are distributed throughout Mercury’s smooth plains (Byrne et al., 2014; Watters et al., 2009), and as these plains have collective ages of 3.9 to 3.5 Gyr (billions of years) as derived from crater statistics (Byrne et al., 2016; Denevi et al., 2013; Fassett et al., 2009; Head et al., 2011; Marchi et al., 2013; Ostrach et al., 2015; Strom, Chapman, Merline, Solomon, & Head, 2008), most if not all of the deformation displayed by those plains must therefore have taken place after about 3.5 Ga (billions of years ago).

The extent to which global contraction, rather than subsidence, deformed the smooth plains is unclear (see section on “Deformation Mechanisms”); however, the structures that deform the more cratered units do not show evidence for having been superposed by those units, and these structures are generally attributed to global contraction (Byrne et al., 2014; Strom et al., 1975). Mercury’s heavily cratered terrains have ages of about 3.9 Gyr to no more than around 4.1 Gyr (Byrne et al., 2016; Marchi et al., 2013; Whitten, Head, Denevi, & Solomon, 2014;), so the structures that deform them are necessarily younger than ~3.9 Gyr. Importantly, the identification with high-resolution image data from MESSENGER of a population of much smaller shortening structures on Mercury—similar to, but between one and two orders of magnitude smaller than, those that deform the cratered plains—indicates that crustal shortening, and probably global contraction, continued into the geologically recent past (Watters et al., 2016; Figure 8). This inference is made on the basis that small-scale structures such as these would likely not survive for billions of years on the Mercurian surface because of the rate of impacts there (Le Feuvre & Wieczorek, 2011; Marchi et al., 2013; Watters et al., 2016).

Figure 8. A small shortening structure (the leading edge of which is marked by white arrows) that appears to crosscut a relatively fresh 500-m-diameter crater (black arrow). This structure is adjacent to, and has the same strike as, Carnegie Rupes (Figure 4B). The image is a portion of MESSENGER image EN1036136378M and is in azimuthal equidistant projection, centered at 59.8°N, 302.7°E. After Byrne et al. (2018).

The age constraints that bracket the timing of crustal shortening on Mercury also help inform us of the timing of extensional deformation. For instance, because Caloris Planitia is about 3.8 to 3.7 Ga (Denevi et al., 2013; Fassett et al., 2009; Strom et al., 2008), the great number of graben must be as well, including those that constitute Pantheon Fossae, which heavily dissected these plains. Such relations hold for the extensional structures within Borealis Planitia (which was emplaced by around 3.7 Ga; Head et al., 2011; Ostrach et al., 2015), as well as those within the Rembrandt basin (the volcanic plains of which date to ~3.7 Ga; Ferrari et al., 2015). Because the volcanic infill within the midsize Rachmaninoff basin may be as young as 1 Gyr (Prockter et al., 2010), then by implication so are the graben therein (Blair et al., 2013). It is difficult to derive dates for the ages of the small graben observed along the crests of some large shortening structures, in part because it is not feasible to acquire areal crater density measurements for landforms of this size and because of fault reactivation. Nonetheless, these extensional features are presumably relatively young for the same reason the small scarps are so adjudged: such graben would likely not survive for long on Mercury’s surface because of the impact bombardment flux there.

As for their origin, the formation timing of the long-wavelength warps on Mercury is unclear. Even so, it is difficult to imagine these warps developing before the emplacement of Caloris or Borealis Planitiae, because the voluminous, low-viscosity lavas there would have presumably coursed around the perimeter of the undulations rather than flowing uphill to completely cover them. But the most convincing evidence for the warps having developed after major plains volcanism is the presence of the craters they have tilted (Balcerski et al., 2012). These craters could only have formed once the lavas cooled and solidified, placing a firm maximum age of the warps at about that of the plains themselves (i.e., ~3.8–3.5 Gyr old; Byrne et al., 2018; Klimczak et al., 2012).

The Influence of Tectonism on Mercurian Volcanism

Chief among the major findings of the MESSENGER mission was the confirmation of major plains volcanism that occurred there (e.g., Byrne et al., 2016; Denevi et al., 2013; Head et al., 2011; Ostrach et al., 2015). However, as discussed previously (see section on “Timing of Deformation”), the majority by area of Mercury’s volcanic smooth plains were emplaced at or by the same approximate time. This record of volcanic activity (i.e., major plains volcanism that presumably pre-dated ~4.1 Ga; (Marchi et al., 2013) that had largely come to an end within the first quarter of the planet’s history (e.g., Byrne et al., 2016)) is markedly different from those of the other inner Solar System planets. Further, many expansive volcanic smooth plains deposits are situated within preexisting impact craters and basins (e.g., Denevi et al., 2013; Fassett et al., 2012; Ferrari et al., 2015; Strom et al., 1975; Figure 3) and the preponderance of sites of explosive volcanism on the planet are collocated with impact or tectonic structures (e.g., Goudge et al., 2014; Jozwiak, Head, & Wilson, 2018; Kerber et al., 2009; Klimczak, Crane, Habermann, & Byrne, 2018; Thomas, Rothery, Conway, & Anand, 2014; Figure 9). Yet these volcanic characteristics are consistent with Mercury’s protracted history of global contraction, evinced by its widespread population of crustal shortening structures (e.g., Solomon, 1977).

Figure 9. An example of a landform interpreted as a pyroclastic vent and surrounding deposit inside the 89-km-diameter Glinka crater. Note that the vent is situated not only in the center of the crater but along the leading edge of shortening structure (shown by white arrows) that crosscuts the crater and extends into the surrounding plains. The figure is part of a three-color mosaic composed of MESSENGER images EW0242128483G, EW0242128487F, and EW0242128491I; colors correspond to red = 996 nm, green = 749 nm, and blue = 433 nm wavelength. The image is in an azimuthal equidistant projection, centered at 11.0°N, 248.0°E. After Byrne et al. (2018).

This consistency arises because magma ascent through the brittle portion of the lithosphere is closely tied to the prevailing tectonic regime (Byrne et al., 2014), which can be characterized as either neutral, extensional, or contractional (Figure 10). The behavior of magma under each of these regimes is controlled by the relative magnitudes of one vertical and two horizontal stress components (with compression again taken to be positive). Under a neutral tectonic regime, the vertical (σV) and horizontal stresses (σH and σh, where σH > σh) are equal to and so governed by the lithostatic stress, σL. Under extensional and contractional regimes, however, the vertical stress corresponds to the overburden stress (i.e., the weight of the overlying rock at any given depth) and tectonic processes act to decrease or increase the horizontal stresses. Magma can migrate in any direction under a neutral tectonic regime (σH = σh = σV = σL) and is therefore equally capable of forming vertical or lateral intrusions (i.e., dikes or sills, respectively), as long as the magma pressure (Pm) reaches the tensile strength of the host rock to actively drive fracturing or remains sufficiently large to act against the lithostatic stress and thus keep preexisting fractures open (Figure 10A). Under an extensional tectonic regime, horizontal compressive stresses (σL = σV > σH > σh) are of sufficiently low magnitude for positively buoyant magma to open (and keep open) vertical fractures for use as conduits through which it can ascend (e.g., Solomon, 1978); the magma, as dikes, can then rise to the surface (Figure 10B). Such a tectonic regime therefore readily promotes magma ascent and effusive and explosive volcanism.

Figure 10. The orientations of magma-filled fractures under different tectonic regimes. Under a neutral tectonic regime (A), vertical and horizontal magma migration is possible as long as the magma pressure (Pm) reaches the tensile strength of the host rocks and remains sufficiently large to act against the lithostatic stress (σL). Under a contractional tectonic regime (B), vertical magma ascent is suppressed, as the vertical stress component is the least compressive and so facilitates the opening of a horizontal fracture when magma pressures reach the strength of the host rocks. Under an extensional tectonic regime (C), magma is capable of ascending vertically, as the least compressive stresses act horizontally and so allow for the opening of vertically oriented fractures. After Byrne et al. (2018).

By contrast, under a contractional tectonic regime, the ascent of positively buoyant magma through the lithosphere is suppressed, as large horizontal compressive stresses prevent the opening of fractures that may otherwise be utilized by dykes (Watanabe, Koyaguchi, & Seno, 1999). Vertical stresses are lower than the horizontal stress components (σH > σh > σV = σL), so horizontal fractures are opened and kept open by magma before stresses can reach the levels necessary for vertical migration. Magma therefore preferentially migrates horizontally under contractional tectonic regimes, producing laccoliths, lopoliths, and sills over the formation of any vertically oriented intrusions (Watanabe et al., 1999; Figure 10C). This stress state dominates during global contraction and has controlled the surface manifestations of volcanism on Mercury.

Asteroid and comet impacts remove overburden, reset prevailing stresses, and either entirely destroy (and thus forms anew) the lithosphere or substantially fracture and weaken the preexisting lithosphere (Byrne et al., 2016), and so it stands to reason that impact sites are the most likely candidates for effusive eruptive activity once global contraction is established. Similarly, that most pyroclastic vents and their associated products are located along or within 20 km of thrust faults or impact structures strongly suggests that such volcanism (or at least that which has been preserved on the surface) was enabled by dense fracture networks along which magma ascended (Klimczak et al., 2018). And observations of thrust faults cross-cutting or otherwise deforming volcanic units on the planet—requiring that at least the most recent increments of faulting occurred after those units were emplaced—supports the overall view of Mercury’s volcanic character and history being strongly moderated by the onset of global contraction.

Summary and Outlook

The Mariner 10 mission revealed Mercury to be a tectonic world—that its outer solid portion has been deformed in response to stress—and MESSENGER broadly helped characterize where, when, and how that deformation took place. We now know that Mercury has experienced a prolonged history of global contraction, wherein its volume has reduced in response to cooling of the interior. This phenomenon has driven horizontal shortening of the planet exterior, producing the populations of scarps and ridges distributed widely across the surface, and which upon its onset inhibited major plains volcanism. Extensional tectonics operated within volcanically flooded impact features, where cooling lavas were largely insulated from the pervasive background compression but were almost entirely absent anywhere else.

This tectonic record contrasts with those of Earth, Venus, and Mars, where major extensional tectonic systems are widespread (e.g., Anderson et al., 2001; Ivanov and Head, 2013) and volcanism has continued without major diminution of scale for billions of years after planetary formation (e.g., Smrekar et al., 2010; Werner, 2009). The tectonism of Mercury most closely resembles that of the Moon, a world that also experienced an extended period of global contraction from interior cooling (e.g., Solomon & Head, 1980). With the discovery of Mercury-sized rocky planets in orbit around other stars (e.g., Barclay et al., 2013), as well as larger extrasolar planets that likely have a similar internal structure (such as the Earth-sized K2-229b; Santerne et al., 2018), it may be that Mercury comes to serve as a template with which to characterize the formation and geological evolution of one-plate planets in general.

Nonetheless, gaps remain in the understanding of Mercury’s tectonic character. For example, it is not clear when global contraction got underway, only that it likely started at some point between 3.8 and 3.5 Ga (e.g., Hauck et al., 2018). Nor is it clear whether tidal spindown has played any meaningful role in the tectonic evolution of the planet. Further, the present-day structure of the lithosphere (including its thermal structure, how the depth of the BDT varies spatially, and crustal thickness) is poorly known, especially in the southern hemisphere where MESSENGER observations were at lower resolutions than those made in the northern hemisphere (Solomon & Anderson, 2018). The discovery of relatively small and well-preserved shortening structures (Watters et al., 2016) suggests tectonic deformation took place in the geologically recent past (and may even be ongoing), but detecting such deformation from orbit is extremely challenging.

Fortunately, the end of the MESSENGER mission does not mark the end of the exploration of Mercury. The joint ESA–JAXA BepiColombo mission (Benkhoff et al., 2010), launched in October 2018 and scheduled to arrive at Mercury in 2025, consists of two orbiter spacecraft designed to operate in parallel. One spacecraft, named Mio, targets the planet’s magnetosphere and exosphere and their interactions with the solar wind. The other spacecraft, the Mercury Planetary Orbiter (MPO), is tasked with investigating the surface and interior properties of Mercury; because its orbit will be less eccentric than that of MESSENGER, the MPO can return higher resolution observations of the southern hemisphere than its predecessor (McNutt, Benkhoff, Fujimoto, & Anderson, 2018).

Of course, the exploration of Mercury should not end with the ambitious BepiColombo mission. Rather, by following the established approach of first flying by, then orbiting, before landing, roving, and ultimately returning samples from an extraterrestrial target (e.g., COMPLEX, 1978), future exploration efforts to the innermost planet should consider the advances possible with landed science and ultimately with samples brought to Earth from the surface. Such endeavors will surely be expensive and technologically complex, but will dramatically enhance the current understanding of the thermal, geological, and tectonic evolution of the enigmatic innermost planet.

Further Reading

  • Blair, D. M., Freed, A. M., Byrne, P. K., Klimczak, C., Prockter, L. M., Ernst, C. M., . . . Zuber, M. T. (2013). The origin of graben and ridges in Rachmaninoff, Raditladi, and Mozart basins, Mercury. Journal of Geophysical Research Planets, 118, 47–58.
  • Byrne, P. K, Klimczak, C, & Şengör, A. M. C. (2018). The tectonic character of Mercury. In S. C. Solomon, L. R. Nittler, & B. J. Anderson, B. J. (Eds.), Mercury: The view after MESSENGER. Cambridge, U.K.: Cambridge University Press.
  • Byrne, P. K., Klimczak, C., Şengör, A. M. C., Solomon, S. C., Watters, T. R., & Hauck, S. A., II (2014). Mercury’s global contraction much greater than earlier estimates. Nature Geoscience, 7, 301–307.
  • Byrne, P. K., Ostrach, L. R., Fassett, C. I., Chapman, C. R., Denevi, B. W., Evans, A. J., . . . Solomon, S. C. (2016). Widespread effusive volcanism on Mercury likely ended by about 3.5 GA. Geophysical Research Letters, 43, 7408–7416.
  • Hauck, S. A., II, Grott, M., Byrne, P. K., Denevi, B. W., Stanley, S., & McCoy, T. J. (2018). Mercury’s global evolution. In S. C. Solomon, L. R. Nittler, & B. J. Anderson (Eds.), Mercury: The view after MESSENGER. Cambridge, U.K.: Cambridge University Press.
  • Klimczak, C., Byrne, P. K., & Solomon, S. C. (2015). A rock-mechanical assessment of Mercury’s global tectonic fabric. Earth and Planetary Science Letters, 416, 82–90.
  • Klimczak, C., Watters, T. R., Ernst, C. M., Freed, A. M., Byrne, P. K., Solomon, S. C., . . . Head, J. W. (2012). Deformation associated with ghost craters and basins in volcanic smooth plains on Mercury: Strain analysis and implications for plains evolution. Journal of Geophysical Research, 117, E00L03.
  • McNutt, R. L., Jr., Benkhoff, J., Fujimoto, M., & Anderson, B. J. (2018). Future missions: Mercury after MESSENGER. In S. C. Solomon, L. R. Nittler, & B. J. Anderson (Eds.), Mercury: The view after MESSENGER. Cambridge, U.K.: Cambridge University Press.
  • Phillips, R. J., Byrne, P. K., James, P. B., Mazarico, E., Neumann, G. A., & Perry, M. E. (2018). Mercury’s crust and lithosphere: Structure and mechanics. In S. C. Solomon, L. R. Nittler, & B. J. Anderson (Eds.), Mercury: The view after MESSENGER. Cambridge, U.K.: Cambridge University Press.
  • Solomon, S. C., & Anderson, B. J. (2018). The MESSENGER mission: Science and implementation overview. In S. C. Solomon, L. R. Nittler, & B. J. Anderson (Eds.), Mercury: The view after MESSENGER. Cambridge, U.K.: Cambridge University Press.
  • Strom, R. G., Trask, N. J., & Guest, J. E. (1975). Tectonism and volcanism on Mercury. Journal of Geophysical Research, 80, 2478–2507.