Map of tectonic shortening structures in Chryse Planitia and Arabia Terra, Mars

ABSTRACT We present a 1:4,000,000 scale map of tectonic landforms in Chryse Planitia and Arabia Terra, on either side of Mars’ dichotomy. Our study area is a ∼3 million km2 region, transitional between Mars’ highlands and lowlands including Oxia Planum, the landing site of the ExoMars rover. Using a structural mapping approach, we digitised all kilometre-scale tectonic structures at a scale of 1:50,000 using high-resolution data (∼6 m/pixel). Although this region is represented as sparsely tectonised on global tectonic maps, we find evidence of widespread tectonic shortening structures across the region. The shortening structures have a dominant N-S orientation and occur in all globally identified geological units. The structural map contributes to a broader understanding of the geological history of the region and Mars’ wider tectonic history.

The tectonic landforms along the dichotomy between the Chryse Planitia lowlands and the Arabia Terra highlands (Figure 1B) have not previously been mapped or quantitively studied using high resolution data.The region comprises predominantly Noachian (∼4.1-3.7 Gyr) and Hesperian (∼3.7-3.0Gyr) terrain (Hartmann & Neukum, 2001;Quantin-Nataf et al., 2021;Tanaka et al., 2014), with a complex geological history and stratigraphy that includes the deposits of extensive river systems (Davis et al., 2016(Davis et al., , 2023;;Fawdon et al., 2022).Oxia Planum (Figure 1B), the intended landing site of the ExoMars rover, is located in this transitional region between the ancient highlands and younger lowland terrains (Vago et al., 2015).High resolution tectonic mapping can provide insights into the timing of tectonic activity in this region, and contribute to a wider understanding of Mars' global tectonic history, and of the evolution and properties of its lithosphere.Additionally, highresolution mapping will provide tectonic context for the ExoMars rover mission.
The impact of tectonism, at both regional and global scale, on the geological history of Chryse Planitia and Arabia Terra is poorly understood.Thus, to provide constraints on the region's tectonic evolution, we present a map of tectonic landforms of a ∼3 million km 2 study area that straddles Mars' global dichotomy (Figure 1; Woodley et al., 2023).Sparse tectonic landforms have previously been identified in the study area; although this region is represented as sparsely tectonised on global tectonic maps (e.g.Anderson et al., 2001Anderson et al., , 2008;;Knapmeyer et al., 2006;Tanaka et al., 2014;Watters, 1993), we identify widespread tectonic landforms termed shortening structures.Shortening structures are tectonic landforms interpreted as the surface expression of thrust-faults with various degrees of folding (including landforms previously described morphologically as wrinkle ridges or lobate scarps; e.g.Golombek & Phillips, 2010).

Location and projection
The map covers an area between 3-45°N and 43-358°W , forming a 1590 km x 1950km rectangle aligned obliquely to follow the dichotomy, and has an area of ∼3 million km 2 (Figure 1B).The map is presented in a Hotine Oblique Mercator projection (azimuth 37°, centre latitude 23°, centre longitude −20°; Hotine, 1946) using the IAU Mars 2000 Sphere.Elevation in the map area ranges between −6450 and 693 m relative to the MOLA datum, and there is a regional slope of <1°dipping north-westward.The southeast of the map area is located in western Arabia Terra, predominantly formed by Noachian highland units (eNh, mNh, lNh;Tanaka et al., 2014).The northwest of the map area is located in eastern Chryse Planitia, comprising Hesperian transition (Htu, eHt, Ht Hto, lHt) and lowland (lHl) units (Tanaka et al., 2014).

Data
We used Mars spacecraft remote sensing data to map kilometre-scale tectonic structures, combining visible, thermal, and topographic data products.We primarily used 793 individual ∼6 m/pixel CTX images and the ∼100 m/pixel THEMIS daytime infrared global mosaic (Edwards et al., 2011) to identify and digitise structures.To aid in structure identification, topographic data was viewed as a semi-transparent layer, including digital elevation models (DEMs) derived from gridded 463 m/pixel (265 pixels per degree) MOLA data and HRSC data (High Resolution Stereo Camera;Jaumann et al., 2007;Neukum et al., 2004).Sampling bias due to solar azimuth for visible data products (potentially under sampling east and west striking structures) is minimised by identifying additional structures using slope maps and using simulated illumination (from azimuths 45°and 315°) products generated from MOLA and HRSC DEMs.
As there is no clear quantitative distinction between morphological categories (Loveless et al., 2023), we opted to use an interpretive structural-based approach to map tectonic features.This is similar to mapping conducted on Mercury (Byrne et al., 2014) where thrust-fault related landforms are grouped under the term 'shortening structures' (see discussion in Byrne et al. (2018) and Klimczak et al. (2019)).We use this structural-based mapping approach, fundamentally relying on the interpretation that these landforms have a tectonic origin.We identify and interpret shortening structures as tectonic, based on: a linear-arcuate planform appearance, an asymmetric profile with a steep limb and a shallow limb (with the exception of asymmetric crenulations), and classic fault-related morphologies and attributes such as postdating crater-rim displacements, en-echelon crenulations, imbricate appearance, and push-up structures (e.g.Fossen, 2010).
We have found some extensional tectonic structures within the study area, but they are too small to be displayed at map publication scale and were therefore not digitised.These include tectonically-induced fractures tens of metres in scale (Apuzzo et al., 2021) and small grabens occurring parasitically on large shortening structures, resembling equivalent examples recently recognised to be widespread on Mercury (Man et al., in press).

Digitisation and errors
We digitised tectonic shortening structures in ArcGIS Pro 2.7.2 software as polylines at a scale of 1:50,000.If there was ambiguity in tectonic origin caused by, for example, proximity to impact ejecta, impact-sculpted terrain, or erosional scarps, the feature was excluded.
We drew the polyline at the intersection between the hanging wall and the footwall, along the break in slope.Following convention, a sawtooth symbol on the polyline records fault-plane dip direction of the inferred underlying thrust fault.This direction is based on the assumption that the steepest limb forms the leading edge/forelimb of the tectonic landform and indicates the orientation of the underlying fault, although we recognise that this is not always the case (e.g.Schultz, 2000).
We calculated the geodesic length of each shortening structure, defined as the trace length in map view, after digitisation (length uncertainty is 2.3% based on repeat digitisation; Woodley et al., 2023).We measured the height, defined as the maximum scarp height of the leading edge, using the gridded MOLA DEM.We generated multiple parallel profiles where relief visually appeared largest, perpendicular to the shortening structure, and recorded the maximum elevation difference as the height (height uncertainty is 5.8% based on repeat digitisation; Woodley et al., 2023).In addition to length and height, the following properties were recorded in the attribute table (Woodley et al., 2023): scarp height class (<50 m, 50-100 m, or >100 m), morphological complexity (scarp-like, broad arch, crenulation, or broad arch and crenulations), degree of erosion (minimally eroded, eroded, heavily eroded), and orientation (linear directional mean as compass orientation).

Results
Widespread tectonic shortening structures occur across the map area.The distribution of the 845 shortening structures is shown in a sample map (Figure 3) and in the Map of Tectonic Shortening Structures in Chryse Planitia and Arabia Terra, Mars (Supplementary Materials; Woodley et al., 2023).
The total length of the 845 digitised shortening structures is ∼28,000 km, with a mean structure length of 33.2 km (range: 4.2-201.8km) and a mean height of 60 m (range: 8-691 m).The shortening structures show large variation in scale, morphological complexity, and degree of erosion.Shortening structures are more abundant in the eastern highland terrain and somewhat less so in the western lowland terrain (Figure 3).Greatest tectonic structure densities occur in four main regions (labelled A -D in Supplementary Materials), three of which are in the highlands (A, C, and D) and one in the lowlands (B).Densities are sparse in the region of Mawrth Vallis, in the northern corner of the study area, and in areas where there are large impact ejecta deposits (Figure 3 and Supplementary Materials).Shortening structures have a dominant N and S striking orientation (N striking structures have underlying fault planes interpreted as dipping eastward and S striking structures have underlying fault planes interpreted as dipping westward), with only few structures striking eastward (NE to SE) or westward (SW to NW; Supplementary Materials).The E/W striking structures are dense in the same areas as the N/S striking structures (Supplementary Materials).Noticeable is the difference in length, with E/W structures being shorter (mean: 24.6 km) than N/S structures (mean: 34.5 km).

Assessment of classification and digitisation method
The interpretive structural-based classification method used for digitisation in this study differs from conventional morphology-based mapping in several ways.Firstly, in the digitisation of linework; a structural-based classification does not distinguish between 'lobate scarps' and 'wrinkle ridges' and both are digitised as the break in slope at the base of the scarp rather than digitising the interpreted fold axis of wrinkle ridges (e.g.Tanaka et al., 2014).Digitising at the base of the scarp (inferred fault trace) is more consistent and objective than digitising fold axes but, at this map's publication scale, the location of fold axes compared to inferred fault traces is effectively coincident.Secondly, the terminology differs as all structures are termed shortening structures instead of being divided based on morphology.Although structures are grouped, their morphological complexity class is recoded in the shapefile attribute table (Woodley et al., 2023).
The morphometry data documenting shortening structures depends on the measurement methods used and thus it is important to consider these.The lengths of the shortening structures (defined as the fault trace length in map view) were calculated as geodesic lengths to prevent distortion due to map projection.However this is a simplified reflection of fault length because: (1) A landform is often a composite of many spatially associated faults; (2) the mappable surface fault trace length is always smaller than the subsurface fault length (e.g.Kim & Sanderson, 2005); (3) small faults are not resolvable at the map scale (the shortest digitised structure is 4.2 km in length); (4) structures are mapped conservatively, which may cause underestimation of their total length, for example when structures are partly covered by impact ejecta a shorter length will be recorded.Consequently, the total mapped length represents the minimum total fault lengths.
The height of shortening structures (defined as the maximum scarp height of the leading edge) was measured by visually identifying where the height appeared largest from the MOLA DEM and slope map, and then drawing multiple MOLA topographic profiles perpendicular to the structure to find the largest scarp height.The maximum heights are based on repeat measurement which are predominantly within ∼10 m of each other.This method was chosen because it was not feasible nor within the scope of the study to make cross-sectional profiles along the entire length of each shortening structure to identify the maximum height.Note that the measured heights are not the 'original formation' heights of the scarps, as variable amounts of erosion have occurred across the study area and have thus lowered the heights.This might include erosion between episodes of fault activity on the same structure.Alternatively, there may have been a net deposition of sediment on the down-dropped side, which would also lessen the height difference.

Comparison to previous structural surveys
Global tectonic trends and orientations on Mars have previously been identified in lower-resolution global surveys (Anderson et al., 2001;Knapmeyer et al., 2006;Tanaka et al., 2014;Watters, 1993) drawing on the 100-300 m/pixel image data sets.In the western hemisphere, thrust-fault related landforms are generally oriented circumferentially to the Tharsis volcano-tectonic province (e.g.Watters, 1993) or are randomly oriented and attributed to global contractional strain (Golombek & Phillips, 2010).The dominant N/S strike of shortening structures in our study area coincides with Tharsis-circumferential orientations.
Shortening structures occur across the study area but are distributed unequally (Supplementary Materials).There appears to be a higher number of structures in the highlands than in the lowlands, with the north and northwest being most sparse.Part of the reason for this is that structures are not clearly discernible in much of the lowlands on visible data products.However, topographic signatures of possible tectonic rises (previously identified by Knapmeyer et al., 2008;Tanaka et al., 2014) are observable underlying a younger textured/fractured mantling deposit in Acidalia Planitia.In contrast to Figure 3, global surveys show a higher number of structures in the Chryse lowlands than the Arabia highlands (Knapmeyer et al., 2006;Tanaka et al., 2014;Tanaka et al., 2005); but this is because they include the mantled tectonic structures in Acidalia Planitia and because the tectonic structures in the highlands are challenging to identify using lower resolution images.

Applications of this study
Studying tectonic features provides insights into the properties and evolution of Mars' lithosphere and its thermal evolution, contributing to our wider understanding of Mars' geological history.Structural maps are invaluable in deducing the tectonic evolution of a region and provide insights into the timings of tectonic activity and sources of stress.The Map of Tectonic Shortening Structures in Chryse Planitia and Arabia Terra, Mars (Supplementary Materials; Woodley et al., 2023) can be used for global, regional, and local scale studies.
For example, the map suggest that tectonic structures in this region are far more common and widespread than previously reported (Anderson et al., 2008;Knapmeyer et al., 2006;Tanaka et al., 2014;Watters, 1993).If tectonic structures are significantly more common globally, then this has implications for various fields of study including the temporal and spatial distribution of seismicity (Knapmeyer et al., 2006), the magnitude of global contraction (Nahm & Schultz, 2011), and for thermal and geodynamic models (e.g.Hauck & Phillips, 2002).The map can also be used to consider the regional contribution of tectonism to topography, and its implications for estimates of palaeoshoreline displacement (Citron et al., 2018).The data recorded in the shapefile attribute table (Woodley et al., 2023) provides a large morphometric dataset of martian shortening structures and could be used to better understand their formation and subsurface structure.If the timing of tectonic activity can be constrained, then tectonic structures become useful stratigraphic markers that can be used to establish the relative ages of crosscutting/crosscut landforms in the area.In future publications, the above applications of the map will be further explored.

Conclusions
A structural map of tectonic shortening structures in Chryse Planitia and Arabia Terra, Mars is presented both as a digital map file map and as a GIS-ready dataset (Supplementary Materials; Woodley et al., 2023).It can be used to aid the understanding of tectonic and geological process along the dichotomy in Chryse Planitia and Arabia Terra.Based on our interpretive structural-based classification and mapping using high resolution data (up to 6 m/pixel), we conclude that: . There are 845 identified thrust-fault related landforms, termed shortening structures, in the study area.We find evidence of extensional tectonic structures, but these are too small to be displayed at map publication scale. .Shortening structures occur across the map area, with a higher structure density in the Arabia Terra highlands than in the Chryse Planitia lowlands. .Shortening structures have dominant north striking (east dipping) and south striking (west dipping) orientations, broadly in line with circum-Tharsis trends (e.g.Tanaka et al., 2014;Watters, 1993). .The structural map shows that tectonic landforms are significantly more common in this region than previously reported (Anderson et al., 2008;Knapmeyer et al., 2006;Tanaka et al., 2014;Watters, 1993).

Software
ArcGIS Pro® 2.7.2 software was used to digitise the data and Adobe InDesign® was used to produce the final map sheet.

Figure 1 .
Figure 1.Location and context of the study area.A) Global setting of the study area (the labelled box) with linework showing previously mapped structures resulting from extensional and compressional tectonism (Tanaka et al., 2014).B) Extent of study area.Oxia Planum is the landing site for the ExoMars rover.Boxes show the location and extent of Figure 2 panels.Mars Orbiter Laser Altimeter (MOLA) hillshade overlain by MOLA digital elevation model.

Figure 2 .
Figure 2. Example of tectonic shortening structures in the map area.THEMIS Day IR images and the location and extent of panels is shown in Figure 1B.Panels show a large variety in appearance, such as large and small structures (A and B) and sinuous and linear structures (C and D).Shortening structures deform impact craters (A and E) and mounds (C and F) and have complex planform patterns (G and H).

Figure 3 .
Figure 3. Sample map of tectonic shortening structures in Chryse Planitia and Arabia Terra.Symbology has been simplifiedsee Map of Tectonic Shortening Structures in Chryse Planitia and Arabia Terra, Mars (Supplementary Materials; Woodley et al., 2023) for full symbology.For location and extent of map, see Figure 1.THEMIS Day IR overlain by MOLA digital elevation model.