Unraveling the Geologic History of Miranda's Inverness Corona

The Voyager 2 spacecraft performed a flyby of Miranda while in the Uranian system in 1986. The Narrow Angle Camera (NAC) (wavelength range 280–640 nm) imaged Miranda at a range of resolutions with eight images at a regional resolution (200–500 m pixel⁻¹). Here, we minimally process the images and mosaic them to create a basemap for our geologic map. We first download the images in Table 1 from USGS Pilot. ⁴ We then process the images using ISIS 3 functions "voy2isis" and "spiceinit." The Voyager 2 images contain reseaux points, so we remove them using the ISIS 3 functions "findrx" and "remrx." After the reseaux are removed, we calibrate the images using the ISIS 3 function "voycal." In order to correct for some aberration within the Voyager 2 NAC, we apply a modulation transfer function (MTF) known to help with the smear in Voyager 2 images (R. Pappalardo 2023, personal communication; archived in the Appendix A with this manuscript) to the image using the ISIS 3 function "kernfilter." Finally, we update the location of each image using the ISIS 3 function "deltack" to refine the alignment of the images in the mosaic. To further align the images, we georeferenced each image in ArcGIS, centering the mosaic on the South Pole as defined by the originally controlled USGS product (Davies et al. 1987; labeled with "SP" in Figure 3).

Zoom In Zoom Out Reset image size

The alignment of the mosaic is not perfect, as can be seen by the small offsets at the image seams (Figure 3). These offsets are particularly noticeable in the images near the limb of Miranda, where the images become more difficult to align using this method. However, our purpose is to map and analyze Inverness Corona, and so the small offsets toward the edge of the mosaic are not of particular concern, nor are they at the scale (>1 km), where it would affect any results.

Zoom In Zoom Out Reset image size

After the completion of the image mosaic, we created a geologic map of the Inverness Corona region (Figure 4). Here, we map geologic units and structures at the finest scale allowable by the resolution of the image. The smallest features that we define as a geologic unit are 5 km across and the smallest features that we map with a line or point are 2 km in width or diameter. We chose these scales to ensure that geologic units were at least 15 pixels across (∼5 km in a 330 m pixel⁻¹ image) and that linear or point features were at least 6 pixels across (∼2 km in a 330 m pixel⁻¹ image). We consider these pixel values the minimum to identify a texture within a unit or the presence of a structure as a linear or point feature.

To gain further insight to the conditions under which Inverness corona formed, we use a previously published DEM of the region to investigate the topographic spacing of features near the northeastern edge of the corona (see Section 2.1 in Beddingfield et al. 2022 for details of DEM generation). The error associated with elevation in this DEM is described in Appendix 1 of Beddingfield et al. (2022) and is estimated to be ±20 m per 1 km. For photoclinometry, the brightness values of each pixel are functions of solar and viewing geometries, and were used to interpret the values of topographic slopes following the methods of Alexandrov & Beyer (2018). The photometric model was taken as a sphere with radii equal to that of Miranda using a Lambertian photometric model. The topographic slopes calculated for each pixel were then incorporated into the DEM. Because albedo variations are visible in this region of Inverness, incorporating multiple Voyager 2 ISS images allowed for differentiation between albedos and slopes for each pixel in the photoclinometry modeling.

Using the DEM, we took profiles perpendicular to a series of scarps in an area that maximizes the length of the profiles within the outer corona region (see Section 4). We use the DEM to take profiles rather than relying on brightness profiles because the brightness variations are not consistently strictly associated with the structures (e.g., some scarps contain bright material whereas others do not). We took the profiles in two regions where there were no apparent craters or other features that would obviously affect the topographic profiles of the outer corona region. We then analyzed the topographic data by taking a Fast Fourier Transform (FFT) of the profiles to agnostically pull out any repeating wavelengths within the data. We performed this on eight profiles across the scarps in the outer corona region. Due to the relatively short signal fed into the FFT and the inherently imprecise nature of repeated structures within the topography (i.e., there is likely some variability around the wavelength of repeated structures), we define a significant peak as one that is at least four times the apparent noise level or a power greater than ∼100. We did find a significant peak, or peaks with a power >100, in the FFTs for each of the topographic profiles. The values and variance in these values across the topographic profiles are discussed in Section 4.

Geologic mapping shows that Inverness Corona consists of an inner chevron-shaped region (inner corona material, ic) made up of a relatively bright material, and an outer region (outer corona material, oc) with a rounded trapezoidal shape made up of material with mottled brightness and linear structures. The outer region structurally consists of lineaments that are quasi-parallel to the outer edges of the terrain with sharp corners or regions of rounded ridge material (r) where the lineaments meet. The boundary of ic is gradational on the edges on the south and west but relatively sharp on the edges to the north and east. The outer boundaries of the oc and the transition to cratered plains material (cp) is sharp with little or no gradation. The transition in morphology of the scarps within oc and directly to the east within cp is particularly interesting. Outside of oc, the scarps are more prominent compared to within oc, which could indicate a shift in deformation mechanism, rheology, or a difference in formation time.

Previous works have observed extensional features within Arden Corona (Greenberg et al. 1991; Pappalardo et al. 1997), and also for Inverness (e.g., Greenberg et al. 1991; Figure 2(c)). Here, we interpret the abundance of scarps and troughs within and around the edges of Inverness Corona as extensional features, which indicate at least bidirectional extension to form the ∼ E-W trending features and the ∼ N-S trending features. The bidirectional nature of the extensional features and angular edges of Inverness Corona is reminiscent of triple junctions on Earth, particularly the Afar triple junction (e.g., McKenzie & Morgan 1969; Mohr 1978).

Another observation from the geologic mapping of Inverness Corona is the identification of bright corona material (bc). We identify several small regions of bc that appear to be made up of clusters of smaller quasi-circular bright splotches. The splotchy texture gives the bc terrain a rough appearance and leads to our hypothesis that this unit is made up of clusters of small craters, perhaps secondaries or pits, that are at or beyond what is resolvable in the current images. If this is the case, then it is interesting that the small craters within the corona are exhuming bright material. A potential layer of bright material has been observed before at the tops of some scarp faces in the region of Verona Rupes (Smith et al. 1986), lending credence to the idea that the small craters could be further exposing this bright layer. The exposure of bright material could indicate that the apparently darker surface of Inverness Corona compared to the surrounding cratered terrain is a relatively thin veneer over brighter material. This is also supported by the presence of a dark material on some scarps and within craters, but not seen as a coherent layer. Alternatively, the relative brightness could be a photometric effect due to rougher material that appears bright in these regions. However, we do not see the relative brightness of this material change consistently with phase angle in the images. Thus, we favor the interpretation that the relatively bright material in these regions is due to a brighter material being exposed from the subsurface, which is indicative of layering in the near subsurface, different compositions, or degrees of weathering.

Immediately outside of Inverness Corona, in the southeast corner, we note a particularly interesting array of scarps that appear almost radially oriented around the southeast corner of the corona. To the west of this arrangement of scarps, along the southern border of Inverness Corona, there are a series of scarps that are subparallel to the southern edge of the corona, similar but less prominent to what is seen on the eastern side of the corona. The radial scarps in the southeast corner are unique in this region due to their apparent orientation. It is unclear if these radial features are associated with the formation of Verona Rupes because this appears to be the termination of this large chasma structure, or if the radial scarps are associated with the formation of the corona itself. Further investigations of the topography in this area could help to differentiate between these two scenarios (e.g., if the scarps were associated with uplifted terrain or had a topographic signature related to that of the corona).

To evaluate the geologic history of the region, we rely on observed cross-cutting relationships between the geologic units. Generally, we find that the cratered plains material, as expected, is the oldest unit on the surface because it is cross cut by all of the other units (Figure 5). Craters (c) appear throughout the stratigraphic column because they cross-cut all other units, and even themselves in a few locations. Crater ejecta (ce) is interpreted to form at the same time as the crater it is associated with. However, because there is nothing that specifically cross-cuts the few craters with identified ejecta, it is unclear if they are the youngest instances that are the only ones to have been retained to present (e.g., due to weathering).

Zoom In Zoom Out Reset image size

The formation of the inner corona material (ic) appears to post-date the formation of the outer corona material (oc) due to the presence of overlapping features in the northeast corner of the ic material. The younger relative age of ic material indicates that the brighter material exposed in ic could be due to age (e.g., not as weathered as oc) or due to the depth of material exposed. There are no deformed craters within the corona material, indicating that the material formed over a relatively short geologic timescale (i.e., before a crater could form and subsequently be deformed). The formation of the rounded ridge material (r) within the oc appears to be either contemporaneous with oc formation or possibly post-dates oc, but likely pre-dates the formation of ic due to the sharp northern boundary of r against ic material.

We interpret the bright corona material (bc, Figure 5) to be younger than the inner and outer corona material because it appears to overprint all of the pre-existing corona terrain. However, because we are uncertain about how the bright corona material forms, it is possible that it has formed over a longer period of time, has subsequently faded (e.g., due to space weathering), and only the brightest spots are currently identified.

Scarp material formed after some crater material (c) and cratered plains (cp), and before other crater material (c) and cratered plains (cp). Interestingly, at least some of the chasma material (ch) and scarp formation post-dates the formation of the outer corona material, which is evidenced by cross-cutting relationships in the northern part of the corona. This is important because Inverness Corona is estimated to be only ∼0.1 Ga (Kirchoff et al. 2022), implying that the formation of at least part of Verona Rupes and the chasma material happened very recently (<0.1 Ga). Notably, this is more recent compared to the expected timing of internal heat sources (e.g., Ćuk et al. 2020; Castillo-Rogez et al. 2023).

The proposed "riser" model would be driven by large-scale upwelling due to diapirism or solid-state convection, driving surface extension (Johnson et al. 1987; Greenberg et al. 1991). The formation of a corona due to a rising diapir would suggest predominantly extensional features, such as the tilt-block style faulting observed in Arden Corona (Pappalardo et al. 1997). While the rising diapir hypothesis would predict a quasi-circular planform for the corona, it has been suggested the pre-existing tectonic fabrics could control the edges of the corona in the presence of radial stresses (Pappalardo 1994). Our observations of Inverness Corona indicate that the region is dominated by extensional features in roughly orthogonal directions, and so would require strong control from pre-existing fractures if the corona was formed due to rising material.

Convection modeling could reproduce the global distribution of coronae (Hammond & Barr 2014), but would require large heat fluxes at least periodically throughout Miranda's history to form the different corona. One hypothesis for driving activity on Miranda is potential past mean motion resonance (MMR) with Umbriel. As suggested in Ćuk et al. (2020), the previous Ariel-Umbriel 5:3 MMR could have spurred significant tidal heating of Miranda, possibly melting its interior and leading to enhanced surface deformation. As such, this past resonance may have formed many of Miranda's younger geologic features, including Arden and Inverness (Ćuk et al. 2020). Elsinore Corona and other geologic features within the cratered terrain appear to predate the predicted Ariel-Umbriel 5:3 MMR, indicating that a more ancient MMR or some other event spurred its formation.

Evidence for the formation of a corona through cryovolcanism (driven by rising material from a diapir) would take the form of presence of flow features, such as lobes of material and/or a vent source, as suggested on Elsinore Corona (Schenk 1991). We do not observe any flow features within Inverness Corona. It is possible that the Voyager 2 images are not resolving the flow features, but potential features related to the flow of material have been observed on Elsinore corona with lower-resolution images, which may suggest that they do not exist on Inverness. The origin of the bright corona material (bc) could be cryovolcanic in origin, but we interpret it as impact ejecta due to the proximity of small craters.

Though not well-supported in recent literature for the formation of Arden or Elsinore (Pappalardo et al. 1997; Beddingfield et al. 2015), we still evaluate the impact and/or reaccretion hypothesis here because it has not been applied to Inverness before. The formation of a corona through impact would likely result in a quasi-circular planform with some degree of radial symmetry, and an associated ejecta blanket. Neither of these features are observed at Inverness Corona. The overall shape of Inverness is closer to an elongated rectangle and contains essentially no radial symmetry. The region around Inverness Corona does appear to be blanketed in regolith that could be interpreted as ejecta (e.g., Beddingfield & Cartwright 2022), but there is no specific association with Inverness because this material seems to be distributed across the cratered terrains and is not present on Arden or Elsinore, which predate Inverness. Associated with the impact hypothesis is the "sinker" model for corona formation (Janes & Melosh 1988), where an impact causes material to sink and form radial or concentric thrusts at the surface, depending on conditions such as the ice-shell thickness. However, we do not observe organized thrust faults within Inverness Corona. It has also been suggested that a large impact caused the total disruption of Miranda and could be responsible for the formation of the corona through the sinking of reaccreted denser material (Janes & Melosh 1988; McKinnon et al. 1991). However, this scenario does not explain the apparent differences in relative ages of the corona (Kirchoff et al. 2022), nor the predominance of extensional tectonic feature observed at the surface.

Of the two aforementioned hypotheses for corona formation, a rising diapir appears to fit the primary observations of the Inverness Corona geologic map—dominated by extension, potential lack of radial symmetry through tectonic control, and relative timing of unit formation. However, we also note that a diapir is not a unique fit to the observations. Other mechanisms, such as ice-shell thickening, could also fit the noted observation. Ice-shell thickening would likely result from a cooling interior and could be driven in the geologically recent past by exiting a recent MMR with Umbriel (Ćuk et al. 2020). Ice-shell thickening could predict large extensional stresses (Nimmo 2004), which when paired with pre-existing tectonic fabrics could create the roughly orthogonal extensional features and trapezoidal planform that we observe. Multiple past episodes of resonance induced heating and then cooling when Miranda exits the resonance could be responsible for the formation of different corona, thus explaining the apparently different ages (Zahnle et al. 2003; Kirchoff et al. 2022). Additionally, there could be other deformed terrain that is coeval with Inverness in the yet unimaged portions of Miranda. For Inverness specifically, we infer that as the ice shell thickened, deformation would transition from distributed to discrete. We see evidence for this transition in the formation of the discrete chasma system of Verona Rupes that evidently post-dates the formation of the more distributed extension seen within Inverness Corona. However, it is important to note that ice-shell thickening does not necessarily explain all of the observations, including the rounded ridges that appear to be indicative of local contraction. Therefore, it is also possible that a stress mechanism such as ice-shell thickening was coincident with another localized stressing mechanism (e.g., a rising diapir) to reproduced all of the observations.

We need more data to further refine the formation mechanism or mechanisms for Miranda's coronae. In particular, imaging with a pixel scale of 50 m and stereo-topography with a vertical precision on the order of 10 m would resolve the fine lineations within Inverness Corona, which would allow a further analysis of their origin (extensional or compressional) and for differentiation between the formation hypotheses for the corona as a whole. Planform morphology alone does not provide a unique interpretation of the geological structures, so stereo-topography coverage of a similar resolution over the entire corona and surrounding structures would provide further constrains on the origin. Additionally, more work should be done to understand if gravity data could resolve a potential mass anomaly (positive or negative) associated with the corona, therefore distinguishing between a rising diapir ("riser," e.g., Pappalardo et al. 1997) formation mechanism and an impact ("sinker," e.g., Janes & Melosh 1988) formation mechanism, if the anomaly is still present. Compositional data that could resolve whether the material exposed within the corona is different than the surrounding terrain would provide evidence for whether the material was sourced in the subsurface, such as through diapirism or cryovolcanism. If Miranda's current ice-shell thickness were known, then it could be compared to estimates derived here and in previous works (e.g., Beddingfield et al. 2015, 2022) of past brittle or elastic layers that are related to the total ice-shell thickness in the past. Finally, because it is quite possible that not all of Miranda's corona have the same formation mechanism, coverage of each corona is critical. Similarly, it would also be important to determine if there are yet unknown coronae on Miranda present in the regions that were not imaged by Voyager 2.