The geography of Oxia Planum

ABSTRACT We present the geography of Oxia Planum, the landing site for the ExoMars 2022 mission. This map provides the planetary science community with a framework to understand this, until recently, unexplored area. The map comprises (1) a mosaic of the panchromatic Context Camera (CTX) Digital Elevation Models (DEM) and Ortho Rectified Images (ORI) controlled to the High Resolution Stereo Camera (HRSC) multiorbit Digital Elevation Models (DEM) and (2) a mosaic of Colour and Stereo Surface Imaging System (CaSSIS) synthetic colour data products, registered to the CTX ORI mosaic. We define a grid of exploration quadrangles (quads) and an informal group of geographic regions to describe Oxia Planum. These regions bridge the scale gap between features observed on large areas (∼100s km2) and the local geography (10s km2) relevant to the Rosalind Franklin rover’s operations in Oxia Planum.


Introduction
ExoMars 2022 is a cooperative mission between the European Space Agency (ESA) and Roscosmos, the Russian space organisation consisting of the ExoMars rover 'Rosalind Franklin' and the instrumented lander 'Kazakoch'. The mission is tasked with searching for signs of past and present life on Mars ( Figure 1) and will investigate the geochemical environment in the shallow subsurface over a nominal mission of 218 Martian days (sols; Vago et al., 2017) in Oxia Planum. The rover's search for signs of life is predicated on identifying the best possible locations for assessing three parameters (1) geological context consistent with life-hospitable conditions, (2) potential chemical biosignatures, and (3) possible physical biosignatures. To select those locations a good understanding of the landing site's geology based on the available orbital information is required.
The first step in the process of identifying the best possible sample locations requires assembling the relevant data and establishing a framework of conventions to describe them. The 'Macro' sub-group of the Rover Science Operations Working Group (RSOWG) was created to characterise the landing site to a fidelity relevant for rover operations; that is, at metre-scale. The rover is capable of driving approximately 30-100 m per day, depending on the terrain's complexity. The RSOWG 'Macrosub-group is studying the landing site (an area on the order of ∼1000 km 2 ) to consider the geological processes that might affect the potential for formation, concentration, and preservation of biomarkers in the 'onesigma' landing ellipse (i.e. the ∼66.75 × 5 km uncertainty ellipse with ∼67% touchdown probability; Figure 1). The 'Macro' sub-group organised a high- resolution geologic mapping campaign between May 2019 and August 2020 (Sefton-Nash et al., 2020). The principal objectives of this exercise were to (i) familiarise scientists with the geography and geology of the landing site and (ii) create a geological map to build and guide the mission's strategic plan, making the best uses of the available resources (e.g. time, power, data) to find locations for sample acquisition and analysis.
As part of this process, several base map data products were created to build a geographic framework in the region around Oxia Planum. We present a map and data package of Oxia Planum to enable the planetary science community to explore the landing site. This includes (1). A grid of exploration quadrangles (quads) and an informal group of regions to describe the geography of Oxia Planum ( Figure 2). These bridge the descriptive scale gap between features named by the International Astronomical Union (IAU) (∼100s km 2 ) and the local landscape (10s km 2 ) relevant to Rosalind Franklin operations. (2). Mosaics of the Context Camera (CTX; Malin et al., 2007) Digital Elevation Models (DEM) at 20 m/ pixel and Orthorectified Images (ORI) at 6 m/ pixel ( Figure 1). These data are projected onto a High Resolution Stereo Camera (HRSC; Neukum et al., 2004) multi-orbit DEM (50 m cell size) and corresponding 12.5 m/pixel panchromatic mosaic (Gwinner et al., 2016), which in turn is fixed to the Mars Orbiter Laser Altimeter (MOLA; Smith et al., 2001) geodetic control network. The CTX mosaic provides a base layer for registration of other higher resolution data sets (e.g. Volat & Quantan-Nataf, 2020). (3). The Colour and Stereo Surface Imaging System (CaSSIS; Thomas et al., 2017) colour data products at 4 m/pixel acquired for Oxia Planum and registered to the CTX ORI mosaic. Data  The CTX camera onboard the Mars Reconnaissance Orbiter provides ∼6 m/pixel panchromatic data with a swath width of ∼20 km (Malin et al., 2007). The main map is a mosaic of six ORI and DEM created from CTX images for which the emission angles and coverage allowed stereo pairs to be used for photogrammetric reconstruction and DEM production ( Figure 1, Table 1). Our CTX DEMs were created following the method of Kirk et al. (2008), using public-domain Integrated Software for Images and Spectrometers (ISIS3) software to pre-process the raw Experimental Data Records (EDR). The EDRs were then processed in SocetSet®, a commercially available photogrammetry suite (http://www.socetset.com), with X, Y, and Z co-ordinates of the DEM controlled to MOLA Point Experimental Data Record (PEDR) data. The ORI and DEM were then post-processed in ISIS, mosaicked in the software Environment for Visualising Images (ENVI), before manual georeferencing in ArcGIS software. Finally, the georeferenced image mosaic was blended in Adobe Photoshop to remove seamlines using the Avenza Geographic Imager extension to retain geospatial information in the blended product.
The output from SocetSet® are 20 m/pixel DEM (resolving topography of ∼50-60 m features) and 12 orthorectified CTX images with a scale of 6 m/pixel. The Expected Vertical Precision (EVP) in each CTX DEM can be estimated based on viewing geometry and pixel scale (Kirk et al., 2003(Kirk et al., , 2008, e.g. EVP = Δ p IFOV / (parallax/height) where Δ p is the Root Mean Square (RMS) stereo matching error in pixel units, assumed to be 0.2 pixels (Cook et al., 1996) and confirmed with matching software for several other planetary image data sets (Howington-Kraus et al., 2002;Kirk et al., 1999). The pixel matching error is influenced by signal-to-noise ratio, scene contrast and differences in illumination between the images. Pattern noise can also be introduced by the automatic terrain extraction algorithm, especially in areas of low correlation. These can be identified as patches of 'triangles' in the hill shade model (e.g. smooth, low contrast slopes and along shadows). IFOV is the instantaneous field of view of the image on the ground (pixel size in metres). If the paired images have different IFOVs, the RMS value is used, e.g. IFOV = √(pixel scale image 1 + pixel scale image 2). The parallax/height ratio, calculated from the three-dimensional intersection geometry, reduces to tan(e) for an image with emission angle 'e' paired with a nadir image, e.g. parallax/height = tan(e), where e = |emission angle 1 − emission angle 2|.

Colour and Stereo Surface Imaging System (CaSSIS)
The Colour and Stereo Surface Imaging System (CaS-SIS; Thomas et al., 2017) instrument on the ESA Trace Gas Orbiter (TGO) continues to observe the ExoMars landing site (Figure 1). CaSSIS collects data with four filters (Infrared (IR); 950 nm, Near-Infrared (NIR); 850 nm, broad transmission, Panchromatic, filter (PAN); 650 nm and BLUE-GREEN; 475 nm), chosen to provide the camera with a limited multispectral capability sensitive to a variety of minerals (Tornabene et al., 2017). CaSSIS has a swath width of ∼9 km and a rotation mechanism to permit stereo acquisitions. We use CaSSIS 3 or 4 band cubes for our scientific investigation of Oxia Planum. A mosaic of synthetic RGB products is presented on the main map. Synthetic RGB products use a combination of PAN and BLUE filter images whereby: The Red channel is the PAN filter mosaic, The Green channel is a combination of a low pass filter of the Blue and a high pass filter of the PAN, incorporating colour information from BLUE and spatial information of PAN. The Blue channel is a combination of PAN and BLUE-GREEN such that each pixel has a value of (2*BLU -0.3*PAN). Each channel is individually contrast-enhanced to form the final product. As TGO operates in a non-sun synchronous orbit, surface overflights repeat every 36 days spanning a range of local times and seasons (Table  2), individual images do not necessarily have appropriate viewing angles, lighting and atmospheric conditions conducive to the creation of a consistent mosaic data set. We will continue to update the database of georeferenced images as more appropriate images are collected by TGO (see Section 6).

Projection
The map is presented in an equirectangular projection centred at, 335.45°E, (24.55°W) based on the IAU Mars2000 sphere. This matches the coordinate system used by the ExoMars Rover Operations Control Centre (ROCC) to minimise local distortion. This is important for maintaining accuracy between remote sensing observation and rover scale operations. And is available with the supporting data sets (Table 4).

Georeferencing and registration
Registration of the CTX DEM mosaic to the HRSC dataset used ∼200 manual tie points between the CTX ORI and HRSC image mosaic and these tie points were then applied to the DEM mosaic. Georeferencing and registration of the CaSSIS data used an initial set of manual tie points to seed the automatic generation of additional tie points using ArcPro 2.7 software. The CTX mosaic and CaSSIS data were rectified using the spline transformation, which optimises for local but not global accuracy (Esri, 2020). This method provided good results for images with a range of viewing angles and accounts well for local adjustments needed for abrupt elevation changes.

Quad grid and contours
The quad grid was created using the ArcPro 2.7 Grid Index Features Tool (Esri, 2021). The grid is a 121 ×

Geographic regions
A common geographical division and naming system for the Oxia Planum region is needed to allow ExoMars team members to communicate efficiently. Identifying and naming geographical locations and zones provides a spatial context for detailed observations, strategic planning and operations, and hypotheses testing.

Differentiating geographic regions
We divide Oxia Planum into 30 regions ( Figure 2 and Regions are smaller closer to the centre of the landing site or where topography and albedo are more variable. This reflects the need to increase the fidelity of discussion where the rover is more likely to land or there are likely to have been more active geomorphic processes. As such these regions capture features pertaining to hypotheses about the paleo-environments being developed by the RSOWG and provide a natural framework to explore Oxia Planum.

Naming geographic regions
The regions were named in three ways: a number, a unique identifier, and a descriptive term. Unique identifiers were drawn from a list of Roman imperial and senatorial provinces at the largest geographic extent of the Roman empire in 117AD. This scheme was chosen because it has geographic and cultural ties throughout Europe and provides an appropriate number and variety of names. The descriptive terms (e.g. Planum, Lacus, etc) are those used in planetary toponomy (IAU, 1979). Names were selected to reflect the geography of the region (e.g. Caledonia has high elevation terrain in the northwest, Aegyptus has a large channel feature). Geographic locations within regions are also named. These names were drawn from a wider list of Roman towns or other relevant geographic locations with suitable, but process-agnostic, descriptive term (e.g. Alexandria Tholus named after the city in the 'Aegyptus' imperial province). These conventions have the capacity to expand this list as an exploration of Oxia Planum continues.
Although IAU recognised features (e.g. Malino crater) have also been included, all other names are informal. Informal naming of local features has been performed by previous Mars Rover mission teams. As has occurred during previous missions, some names will probably be replaced with formal IAU designations as the mission progresses.

Geographic regions
The main map (CTX DEM and ORI mosaics) covers 8750 km 2 and elevations ranging from −2404 m in the South East to −3240 m to the North. The geographic regions and geographic location we identified in Oxia Planum are shown in Figure 2, with further description in Table 3. . Darker toned and 'blueish' surfaces occur in low elevation regions (e.g. Germania), but are also seen capping mesas (e.g. Corsica). The mesa caprocks shed boulders at their marginal scarps, suggesting a consolidated material. However, toward the centre of these regions the colours change to be relatively light toned and 'orangey' (Figure 3) suggesting that this resistant darktoned relatively bluish material is only a thin layer.  High elevation terrain in the far north of the region, forming an arcuate ridge ∼150 m high and 10-20 km across.

08
Cappadocia Planum Plain sloping to the NE formed of a mix of dark and light toned terrains with a promontory in the west of bright layered materials.

Cyprus Craters
Cluster of fresh impact craters and high-standing terrain associated with an infilled and degraded crater rim. 10

Dacia Palus
Low-lying dark-toned terrain with a sinuous southern margin at ∼ −3125 m elevation.

11
Dalmatia Planum Intermediate relief terrain with mottled tones that appears to be transitional between Aquitania Terra and Pannonia Planitia. A small cluster of impact craters with distinct dark ejecta occurs in the west of the zone.

12
Epirus Craters An elongate cluster of small fresh impact craters in the northeast of the region; probably a chain of secondary impact craters originating from a much larger impact far to the southwest.
(Continued) 14 Germania Lacus A low elevation zone of low relief, dark toned plains. Darker toned material dominates towards the edges and the centre has a brighter albedo, bounded to the west by higher elevation terrain and to the east by a subtle moat. Eastern bounding scarp of light toned material with a thin dark top.

15
Hibernia Planum Low elevation terrain in the far north west of the landing site. This region is bisected by a wide, N-S trending ridge. The topography also shows several quasi-circular depressions that are probably ancient impact structures.

16
Italia Promontorium High elevation terrain in the south west. The western part of the region forms the rim of Malino crater. A narrow band of upstanding, terrain connects to Aquitania Terra to the northwest. The rim of Belgica crater is breached in two places, creating a connection between sediment fans to the east of Milano crater.

15849)
Double-layered impact ejecta associated with Kilkhampton crater which lies just outside the map area. Impact ejecta overlies North and South Neocoogoon Vallis in Noricum Promontorium

18
Lugdunensis Planitia Light toned plain in the south of the landing site area that slopes gently to the northeast. A shallow valley in the east (Lutetia Vallis) trends to the north and defines the boundary with Cappadocia to the east. Lugdunensis contains many inverted terrain features, including ridge (Lutetia Dorsum) that runs along the axis of the topographic low containing Lutetia Vallis.
(Continued) Low relief, low elevation plain in the east of the region. The dark surface retains numerous small craters and is brighter at its margins. A scarp of light-toned material, and an external 'moat' define the boundary of the dark-toned plains.

20
Macedonia Promontorium High elevation terrain in the southeast of the region. Light toned layered terrains are overlain by impact ejecta from Kilkhampton crater. The northern boundary is the South Neocoogoon Vallis, a ∼140 m deep and ∼1400 m wide 'U'-shaped valley, that is itself incised into a wider valleyas demonstrated by a terrace along the northern edge of Macedonia Promontorium.

15892)
Malino is a 15 km diameter, degraded impact crater. The floor of the crater is covered by low relief, dark-toned materials that onlap light-toned terrains at breaches in the crater rims to the north and east.

22
Mauretania Terra High elevation, rough, light toned terrain immediately south of the Malino Crater region.

Mesopotamia Palus
Low elevation, light-toned terrain that slopes inwards and to the northeast towards the low-lying region of Dacia Palus, forming a shallow valley.

24
Moesia Planitia Low elevation terrain in the north of the study area. The dark material of the region has sharp contacts with bright surfaces to the south west. Several clusters of bright mounds straddle this margin and are embayed by the dark terrain.
(Continued) Light-toned, layered promontory with patches of superposing dark material. The promontory is incised by North Neocoogoon Vallis and central Neocoogoon Vallis channels.

27
Pannonia Planitia Low lying terrain in the centre of the map area. Several low mesas (Sicilia and Corsica Mensa), each a few kilometres across but only tens of metres high, comprise light-toned layered deposits beneath a thin cap of dark-toned, blocky material.

Raetia Lacus
Low elevation plain comprising a dark-toned surface to the west and a series of flat-topped, light-toned, layered, finger-like ridges to the east. Together, the ridges compose a fan-like landform that is associated with the termination of the North Neocoogoon Vallis. This fan-like feature is interpreted to be a delta. The western topographically low region connects northwards to Aegyptus Vallis and westwards to the Narbonensis Palus.

29
Tarraconesis Terra Low elevation, light-toned plain in the far west of the region. The bright central area has a slightly higher elevation and is strongly associated with clay mineralogy detections .

Thracia Palus
Thracia Pallus is a low elevation zone that forms a several kilometer-wide trough-like valley that extends from Aquitania Terra in the east to Germania Lacus in the West. Remnants of a ridge run down-slope though the centre of the trough.

Impact structures
A variety of impact structures at many stages of degradation occur in Oxia Planum (Roberts et al., 2021 in review at JOM; Figures 3 and 4). The overall physiography of the region suggests a quasi-circular basin. This is bounded to the northwest by the arcuate ridge in Caledonia and to the southeast by the high elevation plateau in Noricum. This could be a crypto impact basin (Figure 3 in Quantin-Nataf et al., 2021) comparable to larger quasi-circular depressions (QCD) in the northern lowland of Mars (Frey, 2006) or stealth QCD (Buczkowski, 2007). Many small impact craters have had their ejecta and rims removed, and most larger impact structures have experienced extensive erosion or inversion. Numerous small QCD, whose origin is unknown but are probably ancient, buried impact craters, are common in the north and central regions of the Oxia Basin ( Figure 3).

Valleys and inverted channels
A variety of channels and valleys occur in Oxia Planum and have complex relationships with the regional geography (Figure 4(a,b)). The largest channels are North, Central and South Neocoogoon Valles in the southeast of the region, which incise Noricum Promontorium and Aegyptus Vallis, which crosscuts Aegyptus Planitia to terminate at dark materials in the low elevation Dacia Palus. The three Neocoogoon Valles are associated with Coogoon Valles, a ∼10 km wide channel system east of the study area (Molina et al., 2017). North and Central Neocoogoon Valles are associated with sediment fans in Raetia Palus that formed after the phyllosilicate-bearing terrains . This leaves an open question about the relationship between the 10 km wide channel as part of the Coogoon Valles and the claybearing terrains in Noricum, Cappadocia, Assyria and Aquitania downslope of it.
Smaller channels (Figure 4(c,d)) occur in the light-toned clay-bearing terrains, and often as low relief valleys terminating in dark terrains (e.g. Mesopotamia). Several of these channels have narrow ridges in the upper reaches (e.g. Lutetia Dorsum) which are in continuum with channels in the lower reaches (e.g. Lutetia Vallis). These regions also have other evidence for an erosional environment, such as periodic bedrock ridges (Favaro et al., 2021;Silvestro et al., 2021), conducive to landscape inversion and inverted crater fill (Roberts et al., 2021 in review at JOM). This context suggests the ridges may be inverted channel deposits (Davis et al., 2016;Pain et al., 2007): ancient river beds exhumed from an alluvial landscape.

. Conclusions
This map and the associated data sets were created to provide a descriptive framework for the geography around Oxia Planum. These data will bridge the gap in scale between regional features (hundreds of square kilometres) and rover operations (hundreds of square metres) and support researchers investigating the Exo-Mars 2022 mission. They will be complemented by larger scale maps (thousands to tens of thousands square kilometres) which will be focussed on regional-scale stratigraphic relationships (Hauber et al., 2021).
From our observations about the geography, we present the following relationships and hypotheses. These serve as part of the ongoing multi-user mapping of the landing site so these data will be explored in future detailed investigations, and could be tested insitu by observations and analyses made by instruments aboard the Rosalind Franklin rover: . The overall topography suggests that Oxia Planum is a crypto impact basin, and the basement is a buried impact structure.

Data availability
The data used in the map, including the informal geographic areas and the Rover Operations Quad grid and multi band CaSSIS cubes being used for the scientific evaluation of Oxia Planum are freely available through the ESA Guest Storage Facility and the Open University Open Research Data Online (ORDO) ( Table 4). A HiRISE orthomosaic and DEM (Quantin-Nataf et al., 2018;Volat & Quantan-Nataf, 2020) was produced for the high-resolution group mapping campaign of Oxia  15N). Darker capping terrains are often light toned towards their centre (arrows) suggesting that the dark material is a thin layer or covering. (D) Bright terrains which correlate with phyllosilicate detections show evidence for layers or layering (as seen in the walls of these two craters) and are clearly seen even in 4 m/pixel CaSSIS data (image; MY35_009394_165_0_RGB; 24.36N, 18.01W) along with period bedrock ridges (Black arrows; e.g. Favaro et al., 2021). Note that the blue colour images represent mafic aeolian materials and the orange/white and yellow toned regions represent the clay-bearing bedrock materials (Parkes Bowen et al., 2021, in review).
Planum (Sefton-Nash et al., 2020). That dataset was produced using the MarsSI infrastructure, is published on the Planetary SUrface Portal (PSUP), and is coregistered with the datasets presented here. We will update this database as new CaSSIS observations are obtained.

Software
The map and other datasets were created and compiled in ESRI ArcPro 2.7. Creation of the CaSSIS RGB products were completed using ISIS3. CTX Digital Elevation Models and mosaicking of the CTX DEM used SocetSet® and Integrated Software for Images and Spectrometers (ISIS3). Georeferencing of CTX mosaic and CaSSIS data were conducted in ESRI ArcPro 2.7. CTX was mosaicked in the software Environment for Visualising Images (ENVI), with seamlines blended in Adobe Photoshop using the Avenza Geographic Imager extension. CaSSIS mosaic was created in ESRI Arc-Pro 2.7.

Open Scholarship
This article has earned the Center for Open Science badges for Open Data and Open Materials through Open Practices Disclosure. The data and materials are openly accessible at and .