1 Introduction

Mars Express began its observations in early January 2004. The targeted duration of the mission was 1 Martian year (687 terrestrial days). However, the mission is still healthy and operating and has been extended until the end of 2026. After almost 20 years in orbit Mars Express remains one of ESA’s most scientifically productive Solar System missions (Martin et al. 2024). Characterization of the surface geology on a local-to-regional scale by Mars Express, ESA’s Trace Gas Orbiter (TGO) and NASA’s Mars Reconnaissance Orbiter allowed constraints to be placed on geological processes in space and time. Stereo and color imaging produced by the Mars Express High Resolution Stereo Camera (HRSC) (Jaumann et al. 2007) with a resolution of 10 m/pixel provided an unprecedented wealth of information about the geology, geomorphology, and chronology of the Martian surface at a regional scale. Despite the planet’s current aridity, past surface water and its geological relevance has also been the major focus of the investigations made by Mars Express. HRSC has sent back thousands of 3D and colour views of the Martian surface at resolution of 10-15 m/pixel, covering almost 90% of the Martian surface. The images show in detail the giant volcanoes and steep-walled canyons, dry river valleys, ancient impact craters and changing polar ice caps. High-resolution digital elevations models on a local to regional scale (∼50 m/pixel horizontal and <10 m vertical resolution) are among the unique strengths of the HRSC instrument. HRSC’s stereo capability also provides basic data for quantitative analyses to constrain the emplacement of volcanic material, fluvial processes, glacial and periglacial surface modification, and aeolian surface/atmosphere interactions (e.g., Neukum et al. 2004; Head et al. 2005; Jaumann et al. 2010; Tirsch et al. 2011; Erkeling et al. 2012; Bamberg et al. 2014; von Paris et al. 2014). Specifically, based on crater size-frequency distribution analyses (e.g., Michael 2013) the chronology of the geological activity could be constrained (see Werner 2009; Jaumann et al. 2015). The paper focuses on the advantage of HRSC to identify the three-dimensional structure of geological features, map geomorphological units and estimate their age.

In addition the OMEGA (Observatoire pour la Minéralogie, l’Eau, les Glaces et l’Activité) (Bibring et al. 2004) imaging spectrometer onboard Mars Express provides comprehensive assessments of surface mineralogy. OMEGA detected hydrated minerals on Mars for the first time. OMEGA data (Bibring et al. 2004) together with HRSC higher resolution and topography data suggested that water played a geologic role on the surface over periods of time. The surface was greatly modified by liquid water-driven processes, which left their signature in extensive surface erosion and depositional landforms. In addition, if water was stable on the surface over extended periods of time, alteration of surface rocks should occur. OMEGA detected phyllosilicates and hydrated sulphates over large, but isolated, areas on the surface (Bibring et al. 2006; Carter et al. 2013). Although both minerals are the result of chemical alteration of rocks, their formation processes are very different. This pointed to periods of different environmental conditions in the history of the planet, and perhaps different source environments (surface and subsurface). The combination of OMEGA spectral maps with HRSC topographic results of higher resolution allow phyllosilicates to be detected at many locations as dark deposits and eroded outcrops at the base of layered sediments. The most water-related geomorphology results come from the high-resolution HRSC topographic results in combination with OMEGA observations, covering steep-walled canyons, dry river valleys, ancient impact craters and changing polar ice caps. The analyses of this combined data base enabled the quantification of geological processes, such as discharges of fluvial valley networks, formation timescales of deltas, sizes, and volumes of sedimentary deposits, as well as estimates of the ages of geological units by crater size–frequency distribution measurements. Both quantifying geological processes and age determinations allow the evolution of Martian water and wind related geologic activities to be constrained. Besides constraining the ages of surface features, HRSC also provides basic data for quantitative analyses to constrain fluvial erosional processes, glacial and periglacial surface modification, and aeolian surface/atmosphere interactions (e.g. Jaumann et al. 2015).

A second major contribution of HRSC is comparative age dating of fluvial, glacial, and lacustrine deposits providing information about the Martian hydrological cycle (e.g. Jaumann et al. 2015).

2 Properties of H2O in Relation to Geological Processes

Water is composed of the two most abundant and chemically reactive elements in the universe – hydrogen and oxygen. Water exists as a liquid at pressures above 6.1 mbar and temperatures between 273 K to 373 K. On planetary surfaces, temperatures are determined by the solar irradiance, which is a function of the solar luminosity (L) and the distance from the Sun (D) as L/D2. In addition, internal heat sources can provide subsurface conditions that allow a liquid phase of H2O in planetary bodies. Water is a highly vital, flexible, changeable, and moveable substance that alters and evolves other materials. The H2O-molecule is a major component of planets, moons, asteroids, and comets and acts in atmospheres, on surfaces and in the subsurface as a weathering agent and a transportation and sedimentation medium. Depending on temperature and pressure, H2O appears as gas, fluid or solid. The H2O-molecule interacts in each physical state with its environment causing varying types of alteration. Water and ice driven erosion is one of the most effective processes for changing solid surfaces.

The hydrogen atom is small compared to oxygen, resulting in an uneven charge distribution of the H2O-molecule. Due to this polar characteristic H2O is able to dissolve and chemically alter almost all natural substances with time. Like most substances, H2O contracts with decreasing temperature. however, below 277 K, due to the beginning of crystallization, it starts to expand and below 273 K it erratically increases its volume by 9%. Therefore, H2O can mechanically destroy almost all natural substances with time. The specific capabilities of H2O to dissolve and break material, makes it a constructor of specific surface and subsurface environments by burying, excavating, and undermining. Freezing reduces the density of H2O with the result that ice floats on water. Thus, bodies of water freeze from top to bottom enabling water to be liquid and active even under extreme conditions. Finally, the low viscosity of water provides high mobility, even for ice, which results in the ability to transport materials and exchange materials between environments. As habitable environments are closely correlated with erosion and sedimentation processes caused by H2O, it is obvious that geological investigations can detect and identify such environments via the tell-tale traces of water and ice.

The versatility of H2O is largest where all its three physical states can exist simultaneously. Earth is obviously such a place, but in the geologic past Mars and the subsurface of some icy moons might also have experienced comparable conditions for water in these three states. In terms of aqueous environments we must distinguish between surface and subsurface conditions. On planets within the habitable zone that are protected by an atmosphere, H2O-driven geologic processes build water and ice bodies within marine, lacustrine, fluvial, and glacial environments. The geologic traces of such processes are shorelines, layered sediments, deltas, river valleys and glacial features such as flow fronts, moraines, eskers, drumlins, and dewatering systems. These water and ice related geologic traces can be found widespread on Mars (e.g. Carr 1996; Masson et al. 2001; Jaumann et al. 2002; Neukum et al. 2004; Head et al. 2005; Jaumann et al. 2005, 2015). Subsurface cavities ranging from sub- μm scale to huge volumes allow water and ice to be present in the underground, protected from harsh surface conditions. Evidence of groundwater and ground ice can be traced on the surface by features such as depressions, retreating erosion, permafrost, thermocarst, and polygons that can also be found on Mars (e.g. Carr 1996; Clifford and Parker 2001; Masson et al. 2001; Jaumann et al. 2002, 2015). In addition, interactions of volcanism or impact bombardment with groundwater and ground ice will trigger hydrothermal processes that can form warm, habitable environments.

3 Fluvial/Lacustrine Related Environments and Activity

Valley networks and rivers: NASA’s Mariner 9 1970s discovered that Mars is heavily incised by troughs that resemble terrestrial fluvial valleys (Masursky 1973). Developments in data resolution has revealed that numerous ancient valley networks dissect the southern highlands of Mars, providing evidence of a past water cycle on the planet (Milton 1973; Carr and Clow 1981; Carr 1996; Hynek and Phillips 2001; Baker 2001; Howard et al. 2005; Hoke and Hynek 2009; Carr and Head 2010; Hynek et al. 2010; Fassett and Head 2008a, 2011; van Berk et al. 2012; Jaumann et al. 2002, 2007, 2014, 2015; Kite 2019).

Most networks formed during the Late Noachian to Early Hesperian period, around 3.7 billion years ago, with surface runoff by precipitation-fed (rain and/or snowmelt) fluvial erosion being the major formation mechanism (e.g., Craddock and Howard 2002; Head and Marchant 2014), with subglacial incision and sapping erosion playing a more minor role (e.g., Grau Galofre et al. 2020a). This period is believed to have been the peak of valley formation (Hynek et al. 2010; Carr and Head 2010; Carr 2012). The formation of valley networks rapidly decreased near the Noachian-Hesperian boundary (Fassett and Head 2008a,b); at later Martian periods (Hynek et al. 2010; Carr 2012), localized incision is observed in some areas (e.g., Dickson et al. 2009; Mangold et al. 2012a; Bishop et al. 2013; Salese et al. 2016; Adeli et al. 2016; Wilson et al. 2016), particularly the flanks of volcanoes (Gulick and Baker 1989; Gulick et al. 1997; Gulick 2001; Basilevsky et al. 2006; Fassett and Head 2007; Hynek et al. 2010; Bahia 2022).

The availability of near global coverage of HRSC images (Neukum et al. 2004; Jaumann et al. 2007), which provides regional scale high-resolution images and allows for the development of higher resolution than previously available regional scale digital elevation models (DEMs) (Scholten et al. 2005; Gwinner et al. 2009, 2010, 2016), has revealed new and vital information about the number, locations, origin, and duration of Martian fluvial activity component.

The Number and Locations of Martian Valley Networks: The most up-to-date global valley map of Mars was mapped using a combination of data from several instruments (Hynek et al. 2010), including THEMIS (Thermal Emission Imaging System) daytime and nighttime IR (infrared) data with a resolution of 100 m/pixel, MOC (Mars Orbital Camera) wide-angle images with a resolution of 231 m/pixel, and MOLA (Mars Orbital Laser Altimeter) topographic data with a resolution of approximately 500 m/pixel. However, mapping efforts using HRSC images has revealed a much greater quantity of valley networks, indicating HRSC can allow for a more comprehensive archive of Martian valley networks (Fig. 1).

Fig. 1
figure 1

HRSC orbit i831 showing extensive dendritic valley networks in the Terra Sabea region. The image is centered at 66° East and 17° South. (DLR/ESA/FU-Berlin)

For example, valley mapping based on HRSC of a latitudinal strip between 20°E–20°W by Bahia et al. (2022) revealed ∼ 1.5 times greater valley drainage density than the previous valley map (Hynek et al. 2010). Mapping of localized areas, e.g., Huygens crater (13.88°S, 55.58°E), the Aeolis region (3.25°S, 219.37°E), and the West Echus plateau (2.61°N, 283.23°E), resulted in drainage densities 3 times greater than those observed in MOLA data (Ansan et al. 2008). While drainage density can be influenced by various factors, such as relief, maturity, and lithology, it is most sensitive to climatic conditions, particularly precipitation intensity and frequency, which affect the kinetic energy of streamflow needed to carve channels and create drainage networks, with high drainage density values generally associated with semiarid regions on Earth that experience intermittent high-intensity rains (e.g, Woodyer and Brookfield 1966). The increase in drainage densities, revealed by HRSC images, indicates that Mars was likely to have been semi-arid during the Late Noachian – Early Hesperian, which is supported by paleolake geometries (Stucky de Quay et al. 2020) and branching angles of valley networks (Seybold et al. 2018). Furthermore, valley mapping using HRSC data has revealed valley networks in locations that they were not previously identified (e.g., west of Hellas Basin (40.29°S, 46.28°E), Libya Montes (1.44°N, 88.22°E)) (Bahia et al. 2022; Jaumann et al. 2010), which holds important implications for inferring the formation mechanisms of valley network from climatic modelling (e.g., Wordsworth et al. 2013, 2015; Bouley et al. 2016; Boatwright and Head 2021, 2022a,b).

The Origins of Martian Valley Networks: HRSC DEMs and images have enabled the high-resolution reconstruction of Martian drainage basins, and in turn, allowed for greater scrutiny of the origins of fluvial activity (e.g., Basilevsky et al. 2006; Fassett and Head 2006; Ansan et al. 2008; Jaumann et al. 2005, 2010; Bishop et al. 2013; Buhler et al. 2014; Grau Galofre et al. 2020a; Boatwright and Head 2022b; Bahia 2022; Bahia et al. 2023). Late Noachian – Early Hesperian valleys are generally attributed to precipitation-fed fluvial activity (e.g., Craddock and Howard 2002). However, Grau Galofre et al. (2020a,b), using the MOLA-HRSC blended DEM and HRSC images, revealed that the profiles and morphologies of Late Noachian – Early Hesperian valley networks indicated the predominant formation mechanisms are fluvial and subglacial erosion, opening the possibility of melting of large-scale ice sheets not only in the southern latitudes (e.g., Head and Pratt 2001; Fastook et al. 2012) but also in the mid-latitudes of ancient Mars with sapping erosion playing a comparatively minor role (e.g. Jaumann et al. 2015). This is supported by Fairén et al. (2012), Hobley et al. (2014), and Buffo et al. (2022), who hypothesize that valleys leading into Gale Crater (5.37°S, 137.81°E) may have formed due to glacial ice-melt. Support for supra-glacial ice-melt as a source for these valley networks was also strengthened by Antarctic terrestrial analogs (e.g., Head and Marchant 2014)

The age analysis of large and sparsely dissected valley networks, such as Nanedi Valles (5.08°N, 49.59°W), Nirgal Vallis (29.10°S, 41.65°W), valleys in fretted terrain, and tributaries of outflow channels and Valles Marineris chasmata (13.75°S, 58.74°W), was conducted using HRSC data sets. Harrison and Grimm (2005) and Jaumann et al. (2015) found that Late Noachian – Early Hesperian valley networks are predominantly of fluvial origin, transitioning to predominantly sapping in origin thereafter. Additionally, Xanthe Terra (1.61°N, 48.04°W) displays older valleys that appear to form precipitation-fed fluvial activity and younger valleys that are sourced from more confined water sources (e.g., groundwater sapping) (Kereszturi 2010, 2014). However, further examination of Naktong Vallis (4.89°N, 33.39°E) revealed a strong link between sapping erosion and overland fluvial activity, suggesting that these processes may have occurred simultaneously in this area (Bouley et al. 2009).

The resolution of HRSC datasets is particularly necessary for the analysis of smaller valley networks, such as those that form during the Late Hesperian – Amazonian periods, revealing that fluvial activity persisted in areas after the Late Hesperian. After the Early Hesperian, Mars’ global climate is hypothesized to be generally extremely arid (e.g., Harrison and Grimm 2005; Bibring et al. 2006), however there appears to be a secondary peak in valley formation during the Early Amazonian (∼2.9 Ga) particularly on the flanks of young volcanoes (Fassett and Head 2007; Hynek et al. 2010). Analysis of HRSC datasets, both morphologically and hydrologically, has shed light on the origins of flank valleys. Most of the small valleys among the flank valleys are attributed to volcanic activity, such as the collapse of lava tubes, followed by fluvial activity (Bahia 2022). On the other hand, larger valleys are mainly caused by ice-melt or snowpack melt due to geothermal heating of high-altitude snow and ice, and not to any major global climate change (Fassett and Head 2007; Bahia 2022). Additionally, very recent (≤25–40 Myr) valley formation on the eastern flank of Olympus Mons (18.34°N, 133.31°W) has been attributed to flowing liquid water induced by dike emplacement leading to cryosphere cracking and groundwater release (Basilevsky et al. 2006).

In other regions of Mars e.g., the Deuteronilus Mensae region (45.11°N, 23.92°E), localized valley formation during the Late Hesperian – Amazonian is largely attributed to ice-melt due to impacts or volcanic activity, or snowpack melting during periods of high-obliquity (Morgan et al. 2009; Dickson et al. 2009; Fassett et al. 2010; Harrison et al. 2010; Parsons et al. 2013; Hobley et al. 2014). Additionally, Peel and Fassett (2013) have observed valleys in 96 Late Hesperian – Amazonian craters, and their formation is commonly attributed to precipitation (likely snowfall) associated with the impact that formed the crater (Peel and Fassett 2013; Goddard et al. 2014).

The Duration of River Incision and Valley Network Formation: HRSC DEMs are more effective in accurately reconstructing Martian drainage basins compared to MOLA DEMs due to their higher lateral resolution when applying paleohydraulic calculations that rely on topographic data such as Hack’s Law and Flint’s Law (Hack 1957; Flint 1974). Using these drainage basins one can gain insight into the duration of fluvial activity required to carve a valley. Hoke et al. (2011) determined that large Martian valley networks have formation timescales indicative of cumulative fluvial activity ranging from 105 to 107 years, with runoff similar to intense streams such as wadis, when flooded in arid regions on Earth. This fluvial activity was likely spread over much greater time periods. For example, examination of valleys incised in the Libya Montes region(1.44°N, 88.22°E) (Fig. 2) reveal that fluvial activity was likely to have extended from ∼4.1 Ga to 3.3 Ga (Erkeling et al. 2011) or even until 1.4 Ga (Jaumann et al. 2010).

Fig. 2
figure 2

a) Oblique view in southern direction showing an extensive valley network in Lybia Montes (HRSC orbit g647). The valley width in the foreground measures 1.5 km. The more longitudinal shaped valley in the center of the image was carved by flowing water and later modified by groundwater-induced processes. The course of the stream is characterized by abrupt changes in direction that indicate the influence of local topography. (HRSC color image superimposed on a HRSC DTM, image bottom to horizon is about 50 km). b) Three-dimensional view of a valley in the Lybia Montes region originating from a lava plane in the south (top), incising an old Noachian massif and entering younger lava plains (middle). At the bottom of the image the channel reenters an older former valley system possibly by a cataract and further discharge to the north. The old channel south and north is older than the channel in the young lava flow indicating different episodes of activity. (HRSC nadir image superimposed on a HRSC DTM, image width is about 60 km).(DLR/ESA/FU-Berlin)

Furthermore, analysis of valley network longitudinal profiles (i.e., elevation vs. valley length) based on HRSC DTM has revealed that the fluvial activity that fueled Late Noachian – Early Hesperian valley formation was likely ephemeral/episodic (e.g., Penido et al. 2013; Buhler et al. 2014; Grau Galofre et al. 2020b).

Localized studies provide evidence that supports the occurrence of episodic fluvial erosion events. For instance, the western Libya Montes valley system and Mangala Valles (11.33°S, 151.38°W) have both shown signs of several such events (Jaumann et al. 2005, 2010). Additionally, research into the intermittency factor of fluvial activity in the Paraná Valles region (23.18°S, 9.79°W) has indicated that valleys were formed by flood events that had intermittent periods of inactivity lasting years to decades (Buhler et al. 2014). Finally, Fastook and Head (2014) explored the distribution of glacial snow and ice in the valley network headwaters predicted by icy highlands climate models (e.g., Wordsworth et al. 2015) and found that the total volume of water estimated to have filled the lakes associated with valley networks could have been supplied by either the raising of surface temperatures to 273 K for a single Mars summer, or a 2000 year period of moderately higher temperatures (warming of +18 K above icy highlands MAT of ∼226 K).

Discharge: River discharge refers to the volume of water that flows through a river channel per unit time. It is typically measured as m3s−1. The discharge of a river can reveal information about the hydrological cycle and the overall water balance of a planet. For example, a high discharge rate indicates that the region received a significant amount of precipitation or had many sources of groundwater or surface water. Discharge values can be calculated from the dimensions (e.g., cross-sectional area and local slope) of Martian channels (e.g., Wilson et al. 2004). To measure these dimensions high-resolution imagery and DEMs are required. Jaumann et al. (2005, 2010) shows that HRSC datasets are sufficient to calculate the discharge values of the interior channel of a 400 km long valley network system located in Libya Montes. They found the discharge of the interior channel was ∼4800 m3s−1, which corresponds to a daily runoff production rate of around 1 cm, which is comparable to terrestrial rivers of the same size (Jaumann et al. 2005). In the absence of HRSC DEMs, the width of a channel (\(W\)) can be used to gain estimates of the bankfull river discharge (\(Q\), the maximum amount of water that a stream or river channel can hold before it overflows its banks and spills onto the surrounding floodplain) (e.g., Irwin et al. 2005). The equation below is for a river that has a bankfull discharge recurrence interval of 2 years:

$$ Q_{bf}\qquad = 0.76(1.9W^{1.22}) $$

Due to preservation issues, interior channels are rare on Mars (e.g., Irwin et al. 2005), however, using HRSC images, Penido et al. (2013) determined that the channels that are present on Mars have a width of around 0.14 of the valley width, meaning that one can make inferences about the discharge that would have been required to form the valley.

Paleaolakes: The same circumstances that led to the creation of valleys on Mars also created favorable conditions for the formation of lakes (e.g., Carr and Head 2010). As catalogued by Fassett and Head (2008b) Mars is home to both open-basin paleolakes (those with an inlet and outlet) and closed-basin paleolakes (those with an inlet and no outlet), the majority of which are Late Noachian – Early Hesperian Lakes. HRSC DEMs were utilized to conduct detailed analyses of the morphology of open basin paleolakes and their corresponding deposits (Goudge et al. 2012a) (Fig. 3).

Fig. 3
figure 3

a) Overview of Jezero crater, derived from the HRSC MC13-east quadrangle mosaic. b) Oblique view of the Jezero delta, view to the west. (DLR/ESA/FU-Berlin)

All the 226 open-basin lakes examined have experienced some level of resurfacing after their formation, due to volcanism (Goudge et al. 2012b), impact crater formation, glacial and periglacial activity and/or aeolian activity. Using HRSC DEMs, Goudge et al. (2015) performed morphometric measurements of 205 closed basin paleolakes, resulting in an inferred maximum volume of water represented by these basins is approximately equivalent to a global equivalent layer (GEL) of water that is 1.2 meters, only a small fraction of the modern water ice reservoir present on Mars (Carr and Head 2015).

The existence of open-basin lakes suggests that certain basins have accumulated enough water to overcome their constraining topography (e.g., Cabrol et al. 1999; Baratti et al. 2015), resulting in disastrous flooding and the carving of outlet canyons (Goudge et al. 2019). Analysis of the outlet canyons with the assistance of HRSC datasets has revealed that despite constituting only ∼3% of the entire length of incised Martian valleys are responsible for ∼24% of the total eroded volume (Goudge et al. 2021). This observation suggests that the canyons were a dominant erosive agent, probably leading to significant modifications in the local topography where they developed (Goudge et al. 2021; Diamant et al. 2022). Some of these outlet canyons display evidence for multiple flow events, indicating that the lakes may have been episodically present (e.g., Balme et al. 2011).

Similarly, paleolakes also display evidence of multiple lake-forming periods. Reconstruction of the hydrological history of an unnamed complex crater in the Xanthe Terra region (1.61°N, 311.96°E) (Di Achille et al. 2006), primarily using HRSC data sets, has revealed that a lake was present during the Late Noachian – Early Hesperian period, with a maximum depth of approximately 700 m, as determined from measurements of a delta-like feature, a central terrace, and two small longitudinal scarps. It is possible that the lake underwent multiple formation periods, and the impact of cratering could have affected the system through heating effects that may have accelerated or rejuvenated the recharge of the local aquifer.

Quantitative measurements of water volumes within Late Noachian – Early Hesperian paleolakes have been calculated using HRSC DEMs. Buhler et al. (2011) examined inlet and outlet valleys, as well as fan morphologies of paleolakes in Erythraea Fossa (27.27°S, 329.06°E) (adjacent to Holden crater (26.04°S,325.98°E) and determined that the three paleolakes were likely precipitation-fed and contained ∼56 km3 of water. In a manner similar to valleys on Mars, the Late Noachian to Early Hesperian lakes are hypothesized to have been supplied by multiple fluid flow events. Based on an analysis of the geometry and morphology of Gale Crater (Palucis et al. 2016), and the depositional fans within it, it is inferred that the at least three principal lake levels were present within Gale Crater that persisted for a duration of at least 1000 years each.

Furthermore, these lacustrine systems formed after the current topographical configuration of Mount Sharp. The Hesperian period (< 3.7 Ga) is also hypothesized to have had similarly large paleolakes. Using HRSC DEMs, Dehouck et al. (2010), examined layered deposits > 300 m thick at the base of a fan within the impact basin Ismenius Cavus (33.89°N, 16.86°E) (an impact basin), revealing that it likely had a 600 m deep lake during this period.

HRSC images used in synergy with HiRISE images have enabled the identification of potential lake strandlines (ridges that are visible markers of past water levels) in Shalbatana Vallis (7.33°N, 317,91°E), indicating that a > 400 m deep lake (Di Achille et al. 2007) was present there during the Hesperian period (∼3.4 Ga) (Di Achille et al. 2009). It is uncertain whether the hydrological activity relied solely on a favorable climate, or if other regional factors, such as volcanism, impact cratering, and tectonism, may have also played a role in initiating or accelerating the local groundwater activity. This suggests that the system may have developed relatively independently of climate conditions.

Late Noachian – Early Hesperian groundwater-fed paleolakes are present in Terra Sirenum (39.49°S, 205.85°E) (Wray et al. 2011). However, it is apparent that other Hesperian paleolakes were possibly climatically sourced, e.g., paleolakes in Ares Vallis (10.29°N, 334.39°E) are hypothesized to be evidence of transient warm and wetter climates (Warner et al. 2010a,b). Salese et al. (2016) utilized hydrological computations of the inflow and overflow rates of the Liberta crater lake (35.23°N, 304.55°E), determining that the entire fluvial network was created within a short to medium (<1000 years) timescale. However, the size and shape of the observed fluvial-lacustrine features imply prolonged periods of activity, akin to similar features found on Earth. The source of water for the 300 km long fluvial system is interpreted to have been derived from melting shallow ice caused by thermal anomalies generated by impact crater formation.

Similarly, the plain encircling Rahway Vallis (unofficially called “Rahway basin”, (8.49°N, 173.57°E)) had a lake that was rapidly filled by fluvial flooding, followed by lake drainage. Possible drainage events may have been instigated by the failure of an ice or debris dam, either during or following the development of the extensive and proximate fluvial flood channel known as Marte Vallis (14.09°N, 182.9°E) (Ramsdale et al. 2015).

The Eridania paleolake system (Adeli et al. 2015; Pajola et al. 2016; Dang et al. 2020), situated along the 180° meridian of Mars, represents one of the most sizable lacustrine environments that existed in the planet’s past. Based on its morphological characteristics, it is believed that this system was comprised of interconnected basins that contained water to a depth of around 2400 m, with a minimum volume of 562,000 km3.

The youngest Amazonian-aged valley network system on Mars (Fassett and Head 2008a), located in association with the ∼225 km diameter Lyot crater in the northern lowlands (Harrison et al. 2010), has been interpreted to be due to the effects of hot Lyot impact ejecta melting of surface snow and ice emplaced during periods of high obliquity (Weiss et al. 2017), and not due to the release of groundwater by the impact (e.g., Russell and Head 2002).

The presence of relatively recent hydrological activity (likely during the Amazonian period (Adeli et al. 2016)) that was likely supported by groundwater replenishment and/or ice/snow melt further corroborates the notion that hydrological activity may have persisted after the Noachian-Hesperian boundary. This is commonly understood to mark the beginning of the current cold and arid Martian climate (Bibring et al. 2006), although Mars is thought to have had a cold climate incapable of sustaining liquid water at the surface after the Late Hesperian.

Sedimentary Deposits of uncertain Origin: While part of the Martian sedimentary record is clearly associated with ancient fluvial or lacustrine environments, a significant portion of it is not clearly associated with a depositional context. These sedimentary deposits of uncertain origin consist of light-toned and layered materials, reaching hundreds of meters to several kilometers in thickness, and mostly occur between ±30° (Tanaka 2000; Malin and Edgett 2000). First identified using Viking Orbiter data, acquisition of HRSC images and digital terrain models permitted improved characterization of these large-scale deposits in terms of their spatial distribution, morphology, and layer attitudes (Fig. 4).

Fig. 4
figure 4

top: HRSC orbit n123. Perspective view on a hydrated sulphate-bearing mound in Tithonium Chasma, Valles Marineris (DLR/ESA/FU-Berlin); bottom: context image (NASA/JPL-Caltech/MOLA; FU Berlin)

These sedimentary deposits are observed in various settings, including in the topographic depressions of the Valles Marineris canyons (or chasmata) (e.g., Mangold et al. 2008; Le Deit et al. 2008; Flahaut et al. 2010; Wendt et al. 2011) the chaotic terrains (e.g., Massé et al. 2008; Sowe 2009; Sowe et al. 2012), impact craters (e.g., Le Deit et al. 2013; Pondrelli et al. 2011a, 2019) and extensive plains such as those in Arabia Terra (e.g., Pondrelli et al. 2015; Schmidt et al. 2021).

In the Valles Marineris, the sedimentary deposits form mounds of material several kilometers thick. Layers are mostly continuous, parallel, and often have a sub-horizontal to shallow dip in the direction of topographic slopes (e.g., Fueten et al. 2008). The combined analysis of HRSC data and the hyperspectral OMEGA data revealed that most of these sedimentary deposits contain mono- and poly-hydrated sulfates, with poly-hydrated sulfates being located topographically higher than monohydrated sulfates (e.g. Gendrin et al. 2005; Mangold et al. 2008; Carter et al. 2024).

Ferric oxides have also been observed locally (e.g., Le Deit et al. 2008; Wendt et al. 2011). Over the years, several depositional environments have been suggested for the sedimentary deposits in the Valles Marineris and other geological settings such as aeolian (Peterson 1981), lacustrine (Nedell et al. 1987; Lucchitta et al. 1992; Quantin et al. 2005), ancient polar deposits made of ice and aeolian dust (Schultz and Lutz 1988), groundwater upwelling leading to spring deposit precipitation (Rossi et al. 2008; Pondrelli et al. 2011a, 2019), the upwelling of hydrothermal fluids in a hypersaline lake (Al-Samir et al. 2017), and interplay between airfall and aeolian processes and groundwater fluctuations (Andrews-Hanna et al. 2007; Mangold et al. 2008).

Overall, the presence of sulfate salts in sedimentary deposits argues for either direct evaporation in a lacustrine environment or cementation of clastic sediments at depth by diagenesis. In addition, these large-scale sedimentary deposits occur in various geological settings, which are likely to have involved mixed origins. Observations of the sedimentary record made in situ by the Mars Exploration Rover Opportunity in the extensive plains of Meridiani Planum and the Mars Science Laboratory Curiosity rover in Gale crater do confirm that several depositional environments are recorded in the hundreds of meters of vertical stratigraphy (e.g., Arvidson et al. 2011; Grotzinger et al. 2015). In terms of the formation age, most of these sedimentary deposits were formed over a long-time span, ranging between the early to late Noachian and possibly as late as the Amazonian (e.g., Loizeau et al. 2012; Tanaka et al. 2014).

Alteration Products: The OMEGA instrument allowed the identification of large kilometer-scale exposures of secondary minerals, mostly clays and sulfates in the Valles Marineris, Mawrth Vallis, Nili Fossae and Sinus Meridiani regions (e.g., Bibring et al. 2006; Poulet et al. 2007, 2008; Loizeau et al. 2007; Gendrin et al. 2005; Mangold et al. 2007; Carter et al. 2023) (Fig. 5).

Fig. 5
figure 5

Top: MOLA map showing the greater Valles Marineris region. All MEX/OMEGA and MRO/CRISM detections of hydrated minerals are represented as green dots (data from Carter et al. 2013). Bottom: HRSC Candor Chasma false color and spectral analyses of different deposits by OMEGA of Candor Chasma (inlet above) (Bibring et al. 2005). ((DLR/ESA/FU-Berlin; NASA/JPL)

This OMEGA global-scale inventory of secondary minerals has been refined and improved by the CRISM instrument, revealing significantly more diverse and widespread aqueous alteration on Noachian and Hesperian terrains than previously seen (e.g., Mustard et al. 2008; Murchie et al. 2009; Carter et al. 2013, 2023, 2024).

Over the last twenty years, spectral observations have been combined with HRSC images and digital elevation models to help constrain the geological context of aqueous mineral deposits. This includes their spatial distribution, morphology, stratigraphy, thickness, and timing of deposition. A significant result of the OMEGA/HRSC joint analyses is the discovery of the most extensive (300 km x 400 km) and thick (≥ 300 m) clay-rich exposure detected on Mars in the region of Mawrth Vallis, formed during the early to late Noachian (e.g., Loizeau et al. 2007, 2010, 2012) (Fig. 6).

Fig. 6
figure 6

Combined OMEGA/CRISM/HRSC mosaic of the Mawrth Vallis region. Fe/Mg-rich materials in red, Al-rich in blue, Sulfate-rich materials in green derived from data from the ExoMars candidate landing site study (Poulet et al. 2020). North is up. (DLR/ESA/FU-Berlin; NASA/JPL)

Several outcrops may contain as much as 50% clay minerals (Poulet et al. 2014). The upper part (<30 m) of the mineralogical stratigraphy consists of Al-clays, such as kaolinite overlying Fe-Mg smectites, which are typically observed associated with weathering profiles on Earth (e.g., Loizeau et al. 2010; Bishop et al. 2008; McKeown et al. 2009). Many extended regions on Mars exhibit weathering profiles, suggesting that pedogenetic alteration of the Noachian basement may have occurred at a global scale during early Mars history, and thus under clement climatic conditions (e.g., Carter et al. 2015; Le Deit et al. 2012; Adeli et al. 2015; Loizeau et al. 2018).

Outflow channels: The giant outflow channels on Mars suggest an origin due to enormous floods of water (e.g. Baker 2001). Measurements of Ares Vallis made with HRSC data contributed to the understanding of the geological history of this outflow channel and helped to constrain the timing of its evolution and to characterize potential climatic variations induced by catastrophic floods (e.g. Jaumann et al. 2015). Ares Vallis extends over 8000 km from about 0.3°N, 342°E to 20.8° N, 324°E, and consists of a system of channels originating from Iani, Hydaspis and Aram Chaos, and debouches into Chryse Planitia. The main valley is divided into two regions characterized by a narrower up-stream reach, about 25 km wide and 1500 m deep, and a broader downstream reach, about 100 km wide and 1000 m deep. Two valley arms, originating from eastern Iani and Hydaspis Chaos, debouch into the main valley. A third valley arm originates from Aram Chaos: it appears as a relatively short, narrow gorge, 10 km wide and 2000 m deep (e.g. Jaumann et al. 2015).

The morphology is characterized by catastrophic flood-induced erosional and depositional features, smaller sinuous channels, and ice-related glacial and periglacial morphologies. Erosional features consist of terraces, streamlined uplands, anabranching channels, giant cataracts, and grooved terrains. Sedimentary features mainly consist of giant bars and pendant bars. Erosional terraces are hundreds of meters high relative to the deepest channel floor, and suggest existence of at least six different, temporally distinct, flooding episodes, based on morphological evidence and impact craters densities (Pacifici 2008; Pacifici et al. 2009; Warner et al. 2009). The catastrophic flow(s) responsible of the outflow channel formation was at least 500 m deep in the narrow reach of Ares Vallis (Pacifici 2008; Pacifici et al. 2009).

High-resolution impact crater statistics acquired from each surface using both CTX and HRSC imagery reveal an Early Hesperian (3.6 Ga) post-flood impact crater population resurfacing age for the topographically highest surface relative to the Early Amazonian post-resurfacing age (2.5 Ga) for the lowest flood terrain (e.g. Jaumann et al. 2015). The HRSC observations suggest an origin mechanism for Ares Vallis and neighboring outflow systems that involved long-lived episodic recharge of water from different sources. These results support more regional outburst models for outflow channel formation (e.g. Jaumann et al. 2015), and render local generation/release of liquid water in a single instance of time less likely. They are also consistent with periodic top-down melting of surface snow and ice and volcano-ice interactions to produce episodic meltwater floods (e.g., Cassanelli et al. 2015; Cassanelli and Head 2019). In addition, ice-related morphologic features of Ares Vallis consist mainly of kame-like features and thermocarst depressions. These features are interpreted as ice-contact deposits by emplacement of sediments in subglacial, englacial or supraglacial ice-walled streams or lakes (Costard and Baker 2001; Pacifici 2008; Pacifici et al. 2009) indicating cool climate condition at least part of formation, an interpretation supported by observations of glacial-like lobes in association with the source of Mangala Valles (Head et al. 2004; Basilevsky et al. 2009).

Oceans: Long before the MEX mission, it was suspected that there might have been an ocean in the northern lowlands of Mars (e.g., Parker et al. 1989; Baker et al. 1991; Parker et al. 1993; Clifford and Parker 1999; Head et al. 1999; Carr and Head 2003). These lowlands, north of the global crustal dichotomy boundary, constitute three-quarters of the planet’s surface area and form the largest watershed on Mars (Smith et al. 1999). Evidence cited in support of the presence of former oceans include its remarkable flatness, the diverse range of distinctive morphological features such as deltas and possible shorelines, the mineralogical composition of specific geologic units as well as the fact that several large outflow channels and countless valley networks drain into the lowlands. These have been interpreted to mean that a significant body of water existed in this area in the past (e.g., Baker et al. 1992; Squyres et al. 1992; Thomas et al. 1992; Fuller et al. 2002; Di Achille and Hynek 2010; Wezel and Baioni 2010; De Blasio 2014; Luo et al. 2017) during bothe the Late Noachian (valley network era) and the Early Hesperian (outflow channel era). However, for the past 20 years, this concept has also been questioned multiple times (e.g., Malin and Edgett 1999; Ghatan and Zimbelman 2006; Head et al. 2018; Sholes et al. 2019; Dickeson and Davis 2020; De Toffoli et al. 2021) and it still represents one of the most significant unknowns in the field of Mars exploration.

Various materials like volcanic deposits, the latitude dependent mantle, glacial and periglacial deposits as well as windblown sand and dust suppress indicative features and obscure spectral observations of the surface that could potentially provide evidence for a former ocean in the northern lowland. As a result, it is difficult to find direct evidence of an ocean within the northern lowlands, and many studies instead focus on landforms along the planetary dichotomy, which is suggested to be the location of at least two ancient shorelines, called the Arabia level/shoreline (Late Noachian) and the Deuteronilus level/shoreline (Early Hesperian) (e.g., Parker et al. 1989, 1993; Clifford and Parker 1999; Head et al. 1999; Carr and Head 2003, 2019).

The utilization of HRSC imagery and topographic data aided in enhancing the understanding of the oceans hypotheses and identifying arguments for and against the ocean theory based on geomorphological and topographical features. Investigations of the southern rim of the Isidis basin suggest that potential coastal cliffs around −3600 and −3700 m represent the Arabia shoreline (Fig. 7) shaped by two distinct stillstands and wave-cut action of a former sea at Isidis Planitia during the Early Hesperian period (Erkeling et al. 2012, 2014). In addition, the authors of these studies interpret the Deuteronilus shoreline to be at −3800 m, which is believed to be the result of the edge of a sublimation residue from a frozen Late-Hesperian/Early-Amazonian sea that may have filled the Isidis basin, similar to the Vastitas Borealis Formation (VBF) in the northern lowlands (Tanaka and Scott 1987; Tanaka et al. 2005; Carr and Head 2003, 2019). Fan-shaped deltaic deposits at Bradbury crater directly adjacent to the supposed Arabia shoreline (Fig. 7) (Erkeling et al. 2012; Bishop et al. 2013; Tirsch et al. 2016, 2018; Bramble et al. 2019) have been interpreted as evidence of past discharge into the proposed Isidis sea (Erkeling et al. 2012).

Fig. 7
figure 7

Perspective view of Bradbury crater at the southern rim of Isidis Planitia. The crater-lake shows a variety of landforms resulting from fluvial, lacustrine, and possibly fluvio-glacial processes. These include degraded valleys, parallel valleys, two cliffs and a terrace, an outlet cut into the northern rim of Bradbury Crater. A delta with Al-rich phyllosilicates, another delta with areas of light-colored, polygonally fractured Fe/Mg phyllosilicates and an alluvial fan are further evidence of the lacustrine activity in this crater. Color-coded HRSC DTM from orbit 2126_0000 on CTX mosaic (from Erkeling et al. 2016)

An ocean in the adjoining Utopia Planitia has also been proposed from an analysis of diagnostic landforms such as mud volcanoes, etched flows, polygons as well as rampart and pancake craters within the plains which suggest a large reservoir of water or mud in central Utopia (Ivanov et al. 2014). Surface observations in Utopia Planitia by the Zhurong Rover have revealed evidence interpreted to support the presence of shallow marine tidal currents in a large body of water in the Utopia basin (e.g., Xiao et al. 2023)

HRSC single strip DEMs and the HRSC-MOLA blended DEM (Fergason et al. 2018) have been used to re-examine the hypothesized shorelines proposed in earlier studies with higher resolution topographic data, resulting in a markedly different picture (Sholes et al. 2019, 2021; Sholes and Rivera-Hernández 2022). These studies argue that the putative coastal landforms lack diagnostic shoreline characteristics, which should, for instance, feature a common elevation level and have a flat top with a step-like flank but are instead curvilinear, tilt in different directions and are laterally not continuous (Sholes et al. 2019).

Moreover, remapping of the global shoreline locations and levels by Sholes et al. (2021) revealed larger differences in elevation as well as lateral offsets of several hundred kilometers. The discrepancies observed between the various levels seem to be attributable to a variety of factors, such as errors resulting from the digitization process, the use of generalizations and extrapolation when determining placement, the combination of data from multiple maps, and the redrawing of sections based on revised interpretations. These findings cast doubt on the efficacy of supposed shorelines as proof of past Martian oceans and suggest the requirement for further meticulous analysis and updated cartography.

HRSC imagery has also been used to discern erosional and depositional features in the northern lowlands that testify to the action of water currents through distinct geomorphology, such as large-scale furrows as well as smaller traction textures (De Blasio 2014). The blended HRSC-MOLA-DEM was used to hypothesize mega-tsunami deposits along the Chryse Planitia margin interpreted to be caused by an impact into a Martian ocean (Rodriguez et al. 2022).

Other studies identify breaks in the longitudinal profile of outflow channel slopes as potential knickpoints that likely formed by base-level changes of an ocean (Duran et al. 2019). These results imply that certain knickpoint elevation zones might have been shaped either during two stable ocean level periods or through multiple phases of fluvial activity combined with varying northern ocean levels (Duran et al. 2019). For instance, the topography of the Kasei Valles outflow channel shows signs of episodic flooding events (ranging from 3.7 to ∼2.0 Ga) interacting with changing ocean/base levels (between 3500 m and 2500 m) in the form of theatre-headed knickpoints that share the same elevation with knickpoints elsewhere on Mars, supporting the hypothesis of a global base-level and thus a global ocean (Duran and Coulthard 2020).

Deltas along the dichotomy boundary are among the most unambiguous geomorphologic landforms indicative of standing bodies of water, and thus have been thoroughly investigated to determine whether they actually attest to an ocean in the northern lowlands. Rivera-Hernández and Palucis (2019) analyzed 50 fan-shaped landforms approximately 500 km east and west of Gale crater and recognized significant elevation differences (up to ∼2400 m) of the delta fronts, which cannot be explained by modelled deformations by Tharsis or polar wander. Thus, the authors conclude that these deltaic deposits are not an indication of a semi-global northern ocean but are rather formed in several larger paleolakes (between ∼330 and ∼13.000 km2 in size) aligned along the dichotomy boundary.

A similar conclusion was drawn from a study that investigated 161 deltas distributed globally to test the hypothesis that these sedimentary formations could delineate an ancient ocean (De Toffoli et al. 2021). The authors analyzed the morphology, elevation, geographic location, and formation time of these fan-shaped deposits and found only six candidates that can potentially be associated with the shoreline of a former Martian ocean. The other deposits show wide discrepancies in age and location and tend to indicate that they were formed in individual smaller standing bodies of water. Thus, relying solely on deltas is insufficient evidence to establish a universally coherent ancient oceanic shoreline. However, deltas can be useful in restricting water levels locally both in terms of space and time (De Toffoli et al. 2021).

4 Glacial/Periglacial Environments and Activity

The Amazonian era on Mars is characterized by ice deposits within the polar caps as well as glacier-related features, latitude-dependent mantles, and remnant debris-covered glaciers in the high to mid-latitudes and at low to tropical latitudes.

Increases in the spin-axis obliquity of Mars cause polar water ice to be mobilized and to be redeposited in the now-colder lower latitude regions (Richardson and Wilson 2002; Haberle et al. 2003; Mischna et al. 2003; Forget et al. 2006) to form regional mantles (e.g., latitude dependent mantles, LDM; Head et al. 2003) and ice sheets, and associated glacial flow features (e.g., lobate debris aprons, LDA; lineated valley fill, LVF; and concentric crater fill, CCF; Neukum et al. 2004; Head et al. 2005; Hauber et al. 2005). When obliquity decreases, the process reverses and equatorial and midlatitude surface ice ablates and returns poleward to be redeposited, leaving some remnant ice deposits protected by a sublimation lag (LDM, LDA, LVF, CCF) or crater ejecta (pedestal craters; Kadish et al. 2010). When mid-to low-latitude glacial ice was not debris-covered, its presence is often revealed by glacial lobe-shaped depressions and related tell-tale features (Head et al. 2008; Hauber et al. 2008).

Head et al. (2003) hypothesized an ice age between 2.1 and 0.4 Ma ago, when the obliquity might have exceeded 30° and Mars was in an “ice age”. During higher obliquity phase, water ice was removed from the polar regions and transported to mid-latitudes. Global circulation models suggest that periods of high obliquity (>30°) may have caused the mobilization of polar volatiles and water vapor in the atmosphere toward the equator and therefore redeposition of ice and snow in lower latitudes (Forget et al. 2006; Madeleine et al. 2009; Mischna et al. 2003). The rapid accumulation of north polar layered deposits resulting from the retreat of ice from the mid-latitudes (Smith et al. 2016) marks the end of this hypothesized ice age. Obliquity variations have been calculated for the past 20 Ma ago (Laskar et al. 2004), but it is likely that the surface of Mars repeatedly experienced such climate changes leading to cycles of deposition and sublimation/evaporation of ice and glacial deposits (e.g. Head et al. 2005; Fassett et al. 2014; Smith et al. 2016). During low obliquity phases, ice and snow are transported back to the poles (Greve et al. 2010; Head et al. 2005; Levrard et al. 2007; Madeleine et al. 2012; Smith et al. 2016).

Mars Express with its HRSC stereo images and high-resolution DEM, greatly contributed to the finding of the major water ice deposits and glaciers on Mars, as well as providing geological context for a deeper understanding of formation and time span of these features.

Past Record: Current Martian ice distribution is largely controlled by insolation and surface thermal characteristics, but in the early periods of Mars’ history a thicker atmosphere would have favored ice deposition both in high elevation and in high latitude regions, as observed on Earth (e.g., Wordsworth et al. 2015; Wordsworth 2016; Palumbo et al. 2018; Kite 2019). Global Climate Models (GCM) of early Mars predict ice deposition in high elevation areas of the highlands, notably in the southern circumpolar region and the equatorial Tharsis bulge. Geological evidence for past ice on Tharsis includes arcuate ridges in Arsia and Pavonis Mons, interpreted to be cold-based moraines deposited by flank glaciers or ice sheets, in agreement with results from Amazonian ice sheet models (Head and Marchant 2003; Parsons et al. 2021; Fastook et al. 2008).

Glacial modification likely also played a role in the sculpture of the outflow channels (Lucchitta 2001, Chapman et al. 2010a and 2010b), and in shaping the basal escarpment and aureole surrounding Olympus Mons, as part of a glaciovolcanic complex (Hodges and Moore 1979; Cassanelli and Head 2019), although submarine or landslide hypotheses have also been raised (De Blasio 2018; Lopes et al. 1980; Tanaka 1985). At the dichotomy boundary, notably around Deuteronilus and Proteronilus Mensae, several landforms are interpreted to be of glacial origin based on analogue landscapes (Fig. 8). This includes U-shaped valleys, cirque-like valley heads, arete-like flank valley walls, and overall canyon scale (Davila et al. 2013).

Fig. 8
figure 8

(a)‘Hourglass craters’ and surroundings as seen by HRSC orbit 0451. Debris covered ice deposits flowed from crater the small crater, converged in the bottleneck between the craters and flowed into the bigger, producing the hourglass morphology. (b) Perspective view to the east the mountainous area is 3000 m high.(DLR/ESA/FU-Berlin)

An informally termed hourglass-shaped crater (Fig. 8) is located in Promethei Terra at the eastern rim of the Hellas impact basin (39° S, 102.8° E).

The impact crater is filled by what was described as a debris-covered piedmont-type glacier based on analysis of HRSC data (Head et al. 2005). The most widespread large-scale past glacial record appears to be in the Noachian-Hesperian southern circumpolar Dorsa Argentea Formation (DAF) (Tanaka and Kolb 2001) and related highland glacial ice accumulations (Wordsworth et al. 2013, 2015, Fastook and Head 2014). The DAF includes sinuous ridges interpreted to be eskers, which formed as sediment accumulated within former subglacial drainage channels (Kargel and Strom 1992; Head and Pratt 2001; Butcher et al. 2016; Scanlon et al. 2018). The record also includes large, elongated basins, interpreted to be subglacial and proglacial meltwater cavities (Head and Pratt 2001), and possibly subglacial volcanoes in the Sisyphi Montes (Ghatan and Head 2002), but does not include characteristic landforms such as drumlins, moraines, or glacial grooves, which are not expected to extensively form under the lower Martian gravity (Grau Galofre et al. 2022).

Current Record: Mars hosts two large polar ice deposits in the north and south polar regions, consisting largely of water ice with a small dust fraction (5-10% respectively, Smith et al. 2020; Orosei et al. 2024). These deposits exchange volatiles and dust with the atmosphere and regolith on timescales from seasonal to hundreds of millions of years, and thus their internal structure holds a long-term record of climate (Byrne 2009; Milkovich et al. 2008, Becerra et al. 2017, 2019). Both the north and south polar deposits consist of a basal unit, a permanent polar layered deposit (PLD), a residual cap, and a seasonal cap (Byrne 2009; Orosei et al. 2024); however, the ages of the two deposits diverge, with the south polar deposit being much older (∼100 Myr) than the northern one (∼10 Myr) (Byrne 2009). Orosei et al. 2024, contains a detailed account of the polar deposits of Mars.

There are various lines of evidence for excess ice in the mid-latitudes of Mars (+/- 30 degrees) (e.g. Bramson et al. 2017; Dundas et al. 2018; Levy et al. 2014; Stuurman et al. 2016; Viola and McEwen 2018; Dundas et al. 2023), both at present and in the past. In addition, water ice at the midlatitude regions has been detected at both shallow (∼1–2 m) and greater depth (a few tens of meters). Fresh impact craters uncover the presence of shallow buried ice in the mid-latitudes (Byrne 2009; Dundas et al. 2018, 2023), and subsurface radar reflections reveal the presence of debris-covered glaciers (Holt et al. 2008; Plaut et al. 2009). However, it is still unclear whether the small-scale debris-covered glaciers on the surface have a similar origin, age, and formation mechanism to the deeper excess ice (Dundas et al. 2018). The formation and distribution of debris-covered glaciers (Fig. 8) are controlled by the climate and therefore provide a unique record of the past and recent climate of Mars (e.g., Forget et al. 2006).

The debris-covered glaciers include lobate debris aprons (LDA) (e.g., Mangold 2003; Hauber et al. 2008), lineated valley fill (LVF) (e.g. Head et al. 2006), concentric crater fill (CCF) (e.g., Levy et al. 2010a), viscous flow features (e.g., Milliken et al. 2003), and glacier-like forms (e.g., Hubbard et al. 2014). Their formation time frame varies among different studies, but they are generally believed to have formed during the Middle to Late Amazonian (e.g., Berman et al. 2015, Head et al. 2005, 2010, Mangold 2003; Parsons and Holt 2016). Tropical Mountain Glaciers (TMG), Latitude dependent mantle (LDM) covers a large part of the mid- and high latitudes in both hemispheres (e.g., Kostama et al. 2006; Mustard et al. 2001; Kreslavsky and Head 2002; Head et al. 2003).

Periglacial Landforms: Geomorphological interpretations of periglacial landforms first aided in the identification of ground ice on Mars from Viking images, with later efforts focusing on the link between distribution and morphology of patterned ground and buried ice, followed by the in-situ detection of a shallow ice table by the Phoenix lander (e.g., Squyres and Carr 1986; Mangold 2005; Morgan et al. 2021).

HRSC data has allowed for a medium resolution characterization of the geomorphology and regional distribution of Martian patterned ground, including stripes, polygons, scalloped depressions, stone circles, mounds, pitted terrain, brain terrain, and lobate gelifluction forms (e.g., Balme et al. 2013), which form due to the near surface thermal expansion and phase changes (melt/sublimation) within ice rich substrates upon daily, seasonal, or obliquity driven surface temperature and pressure conditions. Examples of patterned ground observations focus notably on the northern hemisphere, (e.g., Balme et al. 2013; Korteniemi and Kreslavsky 2013), with some focus on equatorial regions, (e.g., Fairén et al. 2012; Balme and Gallagher 2009), and globally (Levy et al. 2009). Comparisons with terrestrial patterned ground analogues in Svalbard using the DLR airborne HRSC-AX camera, with HRSC-like capabilities (Hauber et al. 2011) have been key to the interpretation of Martian patterned ground types and their implications for the emplacement environment (sublimation/melt) and substrate nature and have highlighted important Earth-Mars differences in scale and morphometry here, as well as in Antarctica (e.g., Levy et al. 2010b).

Multiple processes, including links with ground ice and snowpacks, have been identified as possibly responsible for Martian gully formation. Several studies have concluded that gullies could form or be considerably modified without the need for liquid water, through dry mass wasting, ‘frosted granular flows’ (FGFs), or CO2 sublimation. However, water has also been invoked as a possible mechanism driving gully incision and may involve a combination of several processes, from ice or snowmelt, possibly sourced from latitude dependent mantle, groundwater, or debris flows (e.g., Balme et al. 2006 and Kneissl et al. 2010 for Mars Express derived observations, Harrison et al. 2015 for the most recent global map). Most recently, Dickson et al. (2023) have shown that CO2 mobilized from the south polar region can raise atmospheric pressure sufficiently to permit ice melting and water flow at gulley locations, suggesting a hybrid CO2-liquid water flow origin. See also Thomas et al. 2024.

Climate implication: HRSC, through topographic and image data, has highlighted the glacial landforms that have been key to understanding both the early and recent climate history of Mars (Fig. 9). The presence, direction, and distribution of eskers on the Dorsa Argentea Formation enabled the placing of constraints on the pressure conditions of Mars’ atmosphere at the Noachian-Hesperian transition, as well as its water inventory and surface temperature (e.g. Fastook et al. 2012; Butcher et al. 2017). HRSC observations of LDM and VFF distribution demonstrated the presence of ice on sites where it is not currently stable, implying that past obliquity and atmosphere conditions were once more conducive to the accumulation, crystallization, and flow of ice, and thus that Mars has undergone ice ages (e.g., Head et al. 2003; Head 2011).

Fig. 9
figure 9

Glacial deposits as seen by HRSC. (a) Global map of Mars showing the distribution of ice and glaciogenic deposits on Mars, excluding periglacial terrain and LDM (MOLA/HRSC with hillshade). Late Amazonian north and south polar residual caps (PRC) are drawn in patterned white and blue, whereas Amazonian polar layered deposits (PLD) are shown in dashed blue lines. Hesperian polar units, including the DAf, are shown in light blue in the polar regions. Mid latitude ice is shown in yellow (CCF), blue (LDA), orange (LVF), and younger GLFs are shown in green. Ice rich zones, as determined by the Subsurface Water Ice Mapping (SWIM) program are shown in blue, whereas ice depletion is shown in brown. The four panels show examples of glaciogenic terrains: A CCF (b), an LDA (c), a color-shaded terrain image of Reull Valles LVF (d), a scarp on the NPLD (e) and color-shaded terrain in Ulyxis Rupes showing the SPLD (f) (all north to the top). (b)-(f) (ESA/DLR/FU Berlin)

Once more, the distribution of ice deposits by latitudinal bands with little control by elevation, as is inferred from SWIM team mapping efforts incorporating HRSC data (Fig. 9), implies little adiabatic cooling (thus low atmospheric pressure) since the time of VFF formation. In contrast, findings of tropical glaciers and ice masses on the Tharsis Montes flanks demonstrated fundamental Earth-Mars ice mass balance differences driven by lower surface pressures and low surface temperatures (Fastook et al. 2008). Current presence of ground and buried glacial ice is recorded by periglacial landforms, including polygons and other patterned ground. On Mars, as on Earth, surface and ground ice is fundamental to the understanding of past climate: its distribution records atmospheric and orbital characteristics, its flow the surface and subsurface thermal history, its extent is limited by the water inventory at the time (e.g. Carr and Head 2015), and its presence could provide water sources in recent periods of Mars history.

5 Ages of Erosional/Depositional Activity

The analysis of accumulated crater populations is a key technique in understanding the sequence and timing of action of geological processes on the surface of Mars. Being one of the scientific justifications for the HRSC instrument, the application of the method has considerably expanded over the 20 years of the Mars Express mission. Here we summarise the findings of more than 75 studies of the timing of fluvial, glacial and aeolian processes which made use of HRSC data, often in combination with other datasets.

Among the works analysed, authors used three different Mars chronology systems, commonly referred to as the Neukum, Hartmann and Ivanov chronologies (Hartmann and Neukum 2001; Neukum et al. 2001; Ivanov 2001). Although the systems differ in the detail of their numerical description, they are fundamentally close in concept, even sharing parts of their functional descriptions. While it is possible to convert between model ages derived in the different systems, to do it exactly requires knowledge of the crater populations and size ranges used to make the estimates, which are not always described. We therefore present the model ages as given, regardless of the system. The split timescale as shown on Fig. 10, logarithmic back to 3 Ga and switching to linear thereafter, rather well reflects the time-resolving ability of the crater-counting method: it is robust in distinguishing, say, a third of an order of magnitude below 3 Ga, and likewise in resolving 100 Ma beyond. All three chronology systems should agree to those precisions; for finer differentiation between surfaces, measurements within a single chronology system should be used.

Fig. 10
figure 10

Overview of chronological studies relating to fluvial, glacial and aeolian surface features from crater size–frequency measurements using HRSC data, coloured by geological process (circles: point measurements, bars: intervals, tapered bars: upper/lower limits). See also Jaumann et al. (2015) citations prior to 2015

A summary of the dating measurements of fluvial landforms (in about 70 works), glacial landforms (about 30) and aeolian features (about 8) using HRSC data is shown in Fig. 10. The measurements are shown in groupings for each published work, organised first by type: fluvial (also including other liquid water processes), glacial, or aeolian; then by maximum model age measured in the work. Each entry is accompanied by a brief description of the site and process as well as the citation. Where a publication describes the transition from one type of process to another, these are shown in a single grouping. The Bernhardt and Williams (2021) measurements, exceptionally, are shown over four rows. While the symbols illustrate the spread of measured ages for each category of process, it should be noted that the scale of studied areas can vary greatly between works. The intensity of activity may thus be only loosely related to frequency of measurement.

The transition across the timescale for fluvial features, from valley systems, through basin lakes, sediment and delta deposition, channelling, and surface drainage, up to snow melt and RSL features, reflects not only the timing of occurrence of those particular formation processes in particular locations, but also the bias of our backward-looking view onto the history of the planet. The sequence is simultaneously one of survivability, the evidence of events which can be observed today being only that which has not been obliterated by later activity, which therefore, in general, has to be progressively diminishing in topographic expression. No doubt, many of the events along the timescale were preceded by others of similar type from which traces no longer remain. A symbol of this sort may perhaps be usefully interpreted as a record of the last survivors of traces from different processes, rather than a transparent view of the history of all their activity.

This sequence of survivability relates to a general expectation in crater dating that smaller scale features will be found to be younger. Occasionally proffered as a limitation of the dating technique, that argument is unsatisfying. More directly considered, it results from the fact that as you examine smaller areas, the features you observe relate to a shallower depth of surface material, which is indeed likely to have changed more recently. It is always imprecise to think of a region of surface as having ‘an age’: it is a witness to a complete history of processes which leave traces of different magnitudes. The traces themselves, as structural configurations of surface material, may have ages, and various at different scales can be present at the same location.

Because ice has a lower efficiency than water in moving material, it may be expected to produce features with less topographic expression for comparable time of action, making its traces less survivable, which could be a factor in the later onset of glacial forms. Obviously, for aeolian processes, this efficiency effect is extreme, so that remaining features relate only to recent time periods.

The HRSC studies indicate the availability of liquid water at the surface of Mars through most of its history until recent periods. Modelling of flow transport processes revealed that the formation of deltas on Mars geologically requires only brief timespans (Hauber et al. 2009; Kleinhans et al. 2010) and, based on discharge estimates, the formation of erosional valleys also needs short geological timespans (Jaumann et al. 2005, 2010). A major contribution of Mars Express is the discovery of repeated H2O-related geological processes on Mars (e.g., Neukum et al. 2010). Crater counting model ages (Michael 2013) representing resurfacing events account of the times when fluvial/glacial processes were active, revealed by comparative age dating of, fluvial, glacial, and lacustrine deposits (Neukum et al. 2010), (Fig. 10). Evidence for water related geological activity could be identified from times before 4 Ga until today, showing episodic events with pulses in the intensity of fluvial/glacial processes at ∼3.8–3.3 Ga, 2.0–1.8 Ga, 1.6–1.2 Ga, ∼800–300 Ma, ∼200 Ma, and ∼100 Ma, and possibly a weaker phase around ∼2.5–2.2 Ga ago (Neukum et al. 2010). Some identified geological activities partially coincide with the known age groupings of the Martian meteorites at ∼1.3 Ga, ∼600–300 Ma, and ∼170 Ma (e.g. Neukum et al. 2010; Jaumann et al. 2015).

6 Aeolian Activity and Landforms

The action of wind has left its traces all over the Martian surface. Today, it is the only agent (besides impact erosion) that still shapes the planet’s surface to a notable degree. In its geologic history, Mars provided an abundance of sources and processes (glacial, fluvial, volcanic) that produced fine grained material, that can be transported by wind. In its early history, when the atmosphere was thicker, winds on Mars were even more capable of lifting surface sand and dust particles but also nowadays wind action is still observable. The bedforms and landforms shaped by aeolian processes on Mars are subdivided by the two major types of wind action: deposition and erosion. We focus here on geomorphological features that have been observed and analyzed visually by Mars Express HRSC data. Some features of aeolian deposition and erosion including transverse aeolian ridges (TARs), depositional wind streaks, the north polar erg (Olympia Undae), laminated polar terrain as well as smaller erosional features like ventifacts, grooves, periodic bedrock ridges (PBRs) will not be discussed here.

Depending on the grain size and wind speed and the resulting mode of transport mechanism (suspension, saltation, and creep (e.g., Bagnold 1974)), aeolian deposition results in a dune by saltation or a sand or dust sheet by suspension or creep. The dune type itself is dependent on the degree of sand availability and the wind regime.

Aeolian Deposits

In contrast to Earth, most of the prominent Martian dunes are dark-toned. Lower resolution images (150-300 m/pixel) from Mariner- and Viking-times showed “dark splotches” in impact craters (Christensen 1983; Thomas 1984) associated with emergent wind streaks. But also, widespread dark sand sheets are likewise prominent on the Martian surface and were discovered by telescopic observations long before the age of orbital remote sensing. Already early spectral investigations revealed the mafic composition of the dark aeolian material, dominated by pyroxene and olivine, disclosing its unweathered nature (Carter et al. 2024)

Dune Types and their Distribution: Dune types on Mars comprise all common shapes that occur also on Earth but in different abundances. Their small-scale (< 20 m/pixel) morphology was already identified with visual imagery acquired prior to the MEX mission. The first global inventory and characterization of dunes on Mars was provided by the Mars Global Digital Dune Database (MGD3) – a survey based on THEMIS VIS and MOC NA data and later extended to polar regions and higher resolution CTX data (e.g., Hayward et al. 2007; Hayward 2011; Hayward et al. 2014; Fenton 2020). The statistics revealed that barchan dunes and barchanoid ridges are the most common dune types on Mars (e.g., Breed et al. 1979), followed by transverse dunes, linear dunes, star dunes and domes (e.g. McKee 1979). A non-negligible portion accrues to unclassified dunes, which are either intermediate dune types of unusual morphology or not classifiable due to insufficient image resolution.

A detailed global investigation of dark dunes in Martian craters based on HRSC data has been accomplished (Tirsch 2009; Tirsch et al. 2011). They studied 70 intra-crater deposits across Mars, characterized dune deposition types, mineralogical composition, grain size, formative wind directions and its relation to local wind patterns, determined the potential mobility of the dunes, checked for dependencies of these parameters among each other and with geographical location as well as deduced possible sources of the dark aeolian sediments. Their findings on the dune type distribution agree perfectly with the MGD3 results. In more than half of the 70 studied craters, the aeolian bedforms were associated with sand sheets, which is likewise consistent with the statistics of the MGD3. It is interesting to note the least common type on Mars, i.e., linear, or longitudinal dunes, is the most common type on Earth, whereas barchans dominate the Martian dune types indicating limited sand supply on Mars. Larger dark dune fields on Mars often exhibit a complex dune pattern, with different dune types coexisting or overlapping like in ergs or draas on Earth (Wilson 1972). Silvestro et al. (2008) and Silvestro et al. (2010) analyzed complex dunes fields in the Thaumasia region with HRSC data and found them to comprise barchans, barchanoid ridges, transverse dunes, reversing dunes, star dunes and domes indicating that complex multi-directional wind regimes influenced the development of the dune fields. Secondary dune patterns and slip faces furthermore point to subsequent episodes of aeolian deposition and modification.

Measurements in HRSC DEMs revealed complex dune field heights of up to 400 m and volumes of up to 84 km3 for the ergs in Thaumasia (Silvestro et al. 2008, 2010; Dasgupta et al. 2022). Dune volume calculations by means of HRSC data have also been presented by Jouannic et al. (2012), who studied the Russel crater dune field and estimated a volume of ∼3.9×1010 m3 for this megadune. This number was crucial for the estimation of the possible water content of this frozen dune that in turn was used to explain the occurrence of more than 300 gullies on its slip faces (Reiss and Jaumann 2003).

Chojnacki et al. (2010), Chojnacki et al. (2014a) and Chojnacki et al. (2014b) accomplished comprehensive studies of dune fields in Valles Marineris and used HRSC topographic profiles to evaluate potential sand transport pathways in relation to chasm wall slopes and slip face orientation. The derived chasm slopes showed angles of response that clearly indicate downslope transport of the sand-sized dune material. For many of the Valles Marineris dunes, relatively short (<10 km) and narrow (<1 km) distinct sand transport pathways could be identified by means of HRSC slope profiles (Chojnacki et al. 2014b), proving chasm walls to be local erosional sources for the falling and floor dunes at Valles Marineris.

Dunes and their Relation to Wind Regimes: To analyse whether Martian dunes could be records of past wind regimes and climates a morphological study of 550 dune fields based on HRSC (among other visual imagery) has been conducted by Gardin et al. (2012). The results demonstrated that dunes slip faces are only partially in agreement with current wind regimes predicted by General Circulation Models (GSM) leading to the conclusion that larger parts of them are fossilized and have accumulated under paleoclimate conditions. These findings are consistent with the HRSC-based investigations (Tirsch et al. 2006; Tirsch 2009) who already suspected that dune fields on Mars are constituted of active and fossilized dunes, as indicated by their misalignment with current wind models of the Mars Climate Database (Forget et al. 2007). Some dune fields in this study are even to be indurated or are covered by duricrusts as indicated by their higher thermal inertia in TES and THEMIS data. These key observations and predictions have been recently confirmed by Tianwen-1 Zhurong rover surface observations in the Utopia Planitia region, where dune orientation, cementation and erosion have been linked (Liu et al. 2023) to the end of the recent ice age (Head et al. 2003).

Origin of dark Aeolian Material: A major finding of this global HRSC-based study of dark intra-crater deposits was that re-exposed dark layers (of possible volcanic ash) seem to serve as local sources for the dark aeolian deposits in impact craters on Mars (Tirsch et al. 2011). This conclusion is strengthened by HRSC perspective views showing such layers at crater walls, intra-crater pit walls (Fig. 11), or smaller craters on the floor of larger craters, intersecting dark-toned layers beneath the crater fill.

Fig. 11
figure 11

HRSC perspective view of a dark barchanoid dune field in Rabe crater superposing a large intra-crater pit. Dark layers exposed along the pit walls (white arrows) seem to serve as local sediment sources for the dune material (e.g., Tirsch et al. 2011). The mean diameter of this irregular pit is 55 km. (ESA/DLR/FU Berlin)

Curiously, there is ample evidence of dark aeolian material being transported out of the craters, in the form of wind streaks, but almost no evidence of material being transported into the craters from the outside. HRSC color data suggest that the supposed dark sources in smaller craters are not shadows, but are in fact composed of mafic material, which is typically shown in the blue color in the RGB composites. Confirmation of the mineralogical congruence between the suggested source layers and the dune sands has been provided by spectral analyses of OMEGA and CRISM data presented in this study. Measurements in topographic profiles derived from HRSC DEMs showed that the elevations of craters that expose dark source layers vary in depth, suggesting that multiple layers of buried dark material may exist in the Martian subsurface (Tirsch et al. 2011).

A similar hypothesis was put forward in another HRSC-based study of dark material on the plateaus of Valles Marineris by Chapman et al. (2011), who suggest that the material might have been locally derived from eroded outcrops of lithified dark-toned layered rocks between <10 m and >100 m thick. Central mounds in impact craters in Arabia Terra were analyzed using HRSC and other datasets to answer the question of whether they could serve as local sediment sources (Pan et al. 2019). However, the results did not provide a clear picture, and the authors suggest that young volcanic processes may have contributed to the formation of the mafic sands.

It should be noted that several other sources for the dark dune material on Mars were proposed based on analyses of various orbital datasets. These involve the disintegration of lithified lava (Greeley and Spudis 1981; Edgett and Lancaster 1993), the breakdown of the Medusae Fossae Formation (Burr et al. 2011), the action of glacier retreat at shield volcanoes (Grégoire et al. 2007), erosion of intra-rift sediment sources including interior layered deposits (Chojnacki et al. 2014b) or the erosion of proximal sand-bearing basaltic sedimentary rocks (Ehlmann et al. 2017; Lapotre et al. 2017). For the dunes in Gale crater, different local sediment sources have been suggested. These involve eroded and transported mafic bedrocks from Gale’s crater wall (Ehlmann and Buz 2015) and felsic crater floor materials (Sautter et al. 2015; Lapotre and Rampe 2018).

Deflation is a process by which wind turbulences lifts and removes loose material from the surface. Traction or surface creep, saltation, and suspension are the three mechanisms by which this occurs. Traction or surface creep involves larger grains sliding or rolling across the surface. Saltation involves particles bouncing across the surface for short distances, while suspended particles are fully carried by the wind for long distances.

Abrasion (also termed corrasion) is the process by which suspended or saltated grains wear material from consolidated landforms. Rock clasts shaped by aeolian abrasion develop facets perpendicular to the formative wind, termed ventifacts.

When fine dust particles are transported from the surface into the atmosphere this is known as dust raising or dust lifting. On Mars this is typically caused by wind, as with deflation described above, and thermal changes. Diurnal and seasonal thermal cycles can cause air near the surface of the planet to expand or contract, creating air currents that can lift dust.

Types and Distribution of Erosional Landforms: Aeolian erosion of large bedrock exposures produces yardangs (e.g. Ward 1979; Ward et al. 1985; Melosh 2011; Mandt and Leone 2015) (Fig. 12), which typically exhibit a streamlined shape resembling inverted boat hulls on Earth and Mars. They form parallel to their formational wind in erodible bedrock exposures due to abrasion from entrained sediments.

Fig. 12
figure 12

HRSC orbit 5114 showing extensive yardang fields in the Medusae Fossae formation

They form on highly erodible rocks or consolidated sediment and are on Mars typically observed in the equatorial regions (Ward 1979; Greeley et al. 1993). According to Neukum et al. (2010), yardangs are relatively young features due to their low crater densities. Additionally, Ward (1979) observed that many yardangs do not align parallel to the direction of wind streaks, suggesting that they were formed by different winds from those that created the wind streaks.

HRSC imagery showed that yardangs can exhibit linear ridge and curvilinear morphologies in the Medusae Fossae Formation (Liu et al. 2021). HRSC Digital Elevation Models (DEMs) were used to classify yardang morphology and suggest an evolutionary sequence (Wang et al. 2018).

Wind streaks are linear features that are typically several kilometers long and less than 100 meters wide. Bright wind streaks appear bright and contrast with the darker surface material, while dark wind streaks appear darker than the surrounding surface material. The higher albedo of the bright wind streaks is believed to be due to the removal of darker surface material by deflation, exposing a lighter layer beneath. They are often seen in areas with a lot of loose, fine-grained material, such as sand dunes or rocky terrain. On the other hand, the lower albedo of the dark wind streaks is believed to be due to the accumulation of windblown dust and sand on the surface.

Multitemporal HRSC coverage revealed that both bright and dark wind streaks on Mars are active surface processes can change over time as they are eroded away by the wind or covered over by new deposits of dust and sand. Jehl et al. (2008) demonstrated that multiple HRSC passes over the same region could be used to derive integrated phase functions over a wide range of phase angles, which shed light on the photometric diversity of Martian surface properties. The dark streaks at Gusev crater, the Spirit rover landing site, were deemed to be consistent with large basaltic sand grain sizes organized in more or less packed layers. This was evidenced by their low single scattering albedo, greater backscattering than any other surface across Gusev and Apollinaris, high surface roughness, and present variable surface states. Multi-temporal HRSC and THEMIS-VIS imagery were used to investigate variations in wind streaks and dust devil tracks in Gusev crater, as reported by Greeley et al. (2005) and Greeley et al. (2006). The study revealed that the streaks exhibited a fading trend, resulting in increased brightness, which may be attributed to the deposition of atmospheric dust. Thomas et al. (1981) investigated wind streaks on a northern area of Mars and suggested that they served as a pathway for sand transport within the crater (Rodriguez et al. 2010).

Dust devils (Fig. 13) are low pressure atmospheric vortices that entrain dust and (potentially) sand from Mars’ surface. Much like their terrestrial counterparts, these swirling columns of air are caused when the surface is heated by the sun and the resulting temperature difference between the surface and the overlying air causes a vortex to form. On Earth dust devils are rare events confined to arid environments, whereas Martian dust devils occur at an estimated rate of 1 per sol per km2 (Jackson et al. 2018) ubiquitously across the surface and moreover are three times larger than Earth dust devils (Lorenz 2020). Consequently, they are a major contributor to dust lifting into the atmosphere and the background dust haze (Balme and Greeley 2006), which profoundly influences global climate as these aerosols alter atmospheric thermal structure and circulation (Kahn et al. 1992; Reiss et al. 2011).

Fig. 13
figure 13

Dust devil and dust devil tracks in Peneus Patera as seen in the HRSC nadir image of orbit 2133 (map scale 25 m/pixel). The black square outlines the location of the corresponding SRC image. (b) The higher resolution SRC image (with 5 m/pixel) reveals many more dust devil tracks than seen in the nadir image. (c) A magnified view of the dust devil captured in the SRC image in panel (b), highlighting the kink between its lower and upper sections. North is to the top of each image. (modified from Stanzel et al. 2008)

Before Mars Express, much of these details about the sizes and dynamics of dust devils was unknown, as the only way to detect them was to monitor variations in local meteorological parameters, such as wind velocity, pressure, temperature, and dust loading at landing sites (Ringrose et al. 2003). The nadir and stereo views of HRSC enabled the capture of multiple instances of individual dust devils (Stanzel et al. 2006, 2008), which have been observed to have diameters of 50 meters or greater. The ability to consistently determine the forward motion parameters of dust devils in orbital images has been significantly improved by HRSC, which provides time-sequence imaging of the planet’s surface (Neukum et al. 2004; Jaumann et al. 2007). By analyzing images from the HRSC’s nadir and two stereo channels, Stanzel et al. (2006) were the first to study the motion of dust devils and reported the forward speeds of 205 events across both hemispheres of Mars. They compared the observed forward velocities with the wind regime predicted by the Martian Climate Database (MCD) (Lewis et al. 1999; Forget et al. 1999; Millour et al. 2008) and found that the vortex directions were generally consistent with the modelled winds. However, they also noted cases of chaotic behavior in the dust devil forward directions, such as in Terra Cimmeria. Reiss et al. (2011) compared images from the HRSC, and the Martian Orbital Camera (MOC) taken of Syria Planum with a time delay of up to 12 hours to identify dust devils captured by both instruments. Based on the HRSC images, they estimated an average travel speed of 12 m/s.

7 Climate Record and Implication for Climate History

One of the key parameters for reading and interpreting the record of the ancient climate of Mars is the presence of landforms and deposits that provide proxies for ambient “warm and wet” conditions (evidence for running water, streams, lakes, and oceans; mean annual temperatures (MAT) ≫273K) and proxies for ambient “cold and icy” conditions (polar ice caps, ice-covered surfaces, glacial and periglacial landforms; mean annual temperatures (MAT) ≪273K). A second major proxy for reading the climate record is the nature of the alteration of surface mineralogy (phyllosilicates suggesting warm, water-rich environments; sulphates suggesting the role of volcanic gas exhalations; anhydrous ferric oxides suggest a cold and liquid water-poor climate. A third proxy for reading the climate record is the combination of these two: the geological setting of surface mineralogical alteration.

Mars Express data have provided substantial and compelling evidence for documenting and deciphering the Mars climate record. Although phyllosilicates have been shown to be widespread in Noachian-Early Hesperian environments (Bibring et al. 2006), suggesting a ‘warm and wet’ climate, the geological setting of many of these occurrences are often not in association with other morphological features (valley networks, etc.) indicative of surface water. Some studies have suggested a subsurface alteration source, perhaps related to warm groundwater flow (Ehlmann et al. 2011 Ehlmann and Edwards 2014) and subsequently brought to the surface by impact cratering (e.g., Carter et al. 2013). Others suggest that near-surface weathering under different climate conditions (e.g., Bishop et al. 2018), impact-generated hydrothermal alteration (e.g., Tornabene et al. 2013), torrential hot rains accompanying the Early Noachian large impact basins (e.g., Hellas, Isidis, Argyre; Palumbo and Head 2017), and/or a supercritical steam atmosphere associated with the late stages of the magma ocean era (e.g. Cannon et al. 2017), may have been factors.

The geological evidence for surface liquid water runoff of Mars in the Late Noachian-Early Hesperian is irrefutable; Mars Express data have documented the distribution, topography, morphology, and impact crater size-frequency distribution (CSFD) absolute model ages (AMA) for fluvial valley networks and associated open and closed-basin lakes. These data have also revealed the downslope destination of the combined drainage networks, permitting assessment of their relationship to a hypothesized Noachian northern lowland ocean; recent studies (e.g., Deutsch et al. 2020) fail to find a convincing link. Despite these convincing data, at issue is their relationship to the ambient, or background climate. An excellent case has been made that they represent the presence of a Late Noachian-Early Hesperian ambient ‘warm and wet’ climate, characterized by ambient clement climate with MAT ≫273K. In stark contrast, climate models for this period of Mars history are unable to reproduce such conditions, due to the “faint young Sun” effect (75% current luminosity), even under a wide range of model parameters. Instead, the climate models predict a “cold and icy” climate, with MAT of ∼225K, accompanied by an adiabatic cooling effect that produces snow and ice deposits in the high topography above about +1 km elevation. Although this ambient ‘cold and icy highlands’ scenario places glacial water as ice at the headlands of many valley network systems, it requires a transient heating mechanism (e.g., release of greenhouse gases, etc.) to melt the snow and ice to form the valley networks, open and closed-basin lakes. This debate is currently ongoing and Mars Express data are being used to marshal evidence to test both hypotheses, and to propose new experiments and future missions.

Geomorphic evidence for huge outflow channels, sourced primarily in the circum-Tharsis region, and mineralogic and geomorphic evidence for Interior Layered Deposits (ILD) on the floor of Valles Marineris, are also proxies for ambient climate conditions in the Hesperian. What is the origin, state, and depositional environment of the water that produced the outflow channels? Most hypotheses call on catastrophic release of overpressurized groundwater for the outflow channels, an interpretation that requires a global cold and icy ambient climate and a continuous global cryosphere that is sufficiently thick in equatorial regions. On the other hand, processes for the closely contemporaneous ILD deposits and adjacent similar deposits at Meridiani, require equatorial rainfall, shallow groundwater migration, and groundwater upwelling at mid-latitudes, a climate interpretation inconsistent with the presence of a global cryosphere. Furthermore, in a global cold climate, Tharsis is predicted to be the site of extensive snow and ice deposits, and the confluence of extensive Tharsis volcanism and surface snow and ice have compelled researchers to re-examine the role of volcano-ice interactions as a potential source of large quantities of liquid water. Whatever its origin, what is the fate of the massive volumes of water that flowed through the outflow channels? Detailed analysis of morphology and ages from Mars Express data show that the water debouched into the northern lowlands, and many investigators believe that it filled the lowlands with a vast ocean of liquid water that persisted for significantly long periods of time, such that subsequent impacting bolides into the ocean produced massive tsunamis (De Blasio 2020). Others have postulated that the outflow channels formed episodically, each separated by tens to hundreds of millions of years, and that any liquid water flowing into the northern lowlands would rapidly freeze and sublimate, returning to global cold traps. Mars Express data are essential to addressing these conundra, and in designing and planning new missions to help resolve them.

Mars Express data for Amazonian-aged morphologic features, deposits, and mineralogy have provided widely accepted evidence that the ambient Mars climate was cold, dry, and icy. Surfaces of these ages are characterized by Antarctic Dry Valley-like alteration environments (anhydrous ferric oxides), typical of very cold ambient temperatures, low atmospheric water abundances, and generally anhydrous conditions. Morphologic and topographic data and investigations from Mars Express have amply and convincingly compiled copious amounts of supporting data for the presence of this ambient climate in the form of a wide variety of glacial and periglacial landforms (e.g., lobate debris aprons, lineated valley fill, glacial mantles, contraction-crack polygons, etc.). Located at a full range of latitudes (all the way down to tropical mountain glaciers), these features, and the CSFD absolute model ages enabled by Mars Express data, reveal a record of the importance of obliquity changes driving mobilization of polar ice and its redistribution to varying latitudes throughout the Amazonian, as a modulator of the ambient cold and icy climate. The presence of gullies (alcove, channel, and fan) has been interpreted to indicate melting of snow and ice even under these extremely cold and dry conditions, similar to today’s Antarctic Dry Valley climate, but alternate “dry” scenarios involving CO2 sublimation appear to be a convincing contributor to currently observed changes in the gully morphology. Current debate centers on whether one process or the other dominates, or whether they might operate in concert over longer timescales.

Taken together, Mars Express geomorphological data have provided a giant leap in our documentation and understanding of the record of water-related process on Mars, and as proxies for deconvolving the complex and debated climate history of Mars. In addition to a significant increase in our understanding, these Mars Express data have sharpened the new questions that have arisen, giving rise to many concepts and future mission proposals to address outstanding question and resolve remaining conundra.

8 Summary and Conclusions

Mars Express data have provided substantial insights into the geological record of water-related processes on Mars, documenting the morphological characteristics, topographic relationships, and chronology of the associated landforms and deposits. Together, these data have revealed a consistent and compelling picture of the role of liquid water flow during the Late Noachian-Early Hesperian (e.g., valley networks, open and closed-basin lakes, and possible northern lowland oceans) and the role of these features in distinguishing between ambient “warm and wet” and “cold and icy” climates that might have been characteristic of early Mars. For the Early Hesperian period, Mars Express data have helped in characterizing volcanic landforms and their chronology, revealing evidence for volcano-ice interactions and the generation of meltwater, the role of volcanic eruption-induced warming and ice melting, as well as the potential contributions of degassed water to the atmosphere and its effects on climate. Important in this regard is the detailed morphologic, mineralogic and topographic characterization of the ILDs, dominated by sulphate deposits in various hydration states. For the Late Hesperian and Early Amazonian period, Mars Express data have revealed details of the morphology and chronology of the massive outflow channels, enabling improved assessment of the quantity of liquid water involved, the timing of its emplacement, and the testing of hypotheses for the formation and fate of a northern lowlands’ ocean. For the Amazonian period, Mars Express data have enabled discovery and documentation of numerous occurrences of non-polar ice deposits, including tropical mountain glaciers associated with the Tharsis Montes; taken together, these deposits have been dated and reveal the presence of ancient ice preserved beneath sublimation till, providing a history of obliquity-induced climate change, and a roadmap for future robotic and human exploration missions. Gullies (alcoves, channels, and fans) mapped in association with a geologically recent obliquity-induced latitude-dependent ice mantle, have suggested either meltwater or CO2-assisted erosion as the dominant process, a debate that continues today. All these features provide examples of various environments that are known to harbor life on Earth. These discoveries and hypotheses thus provide numerous compelling questions that can be address with future orbital-lander-rover-sample return robotic missions and coordinated human exploration expeditions.

Geomorphological features of aeolian deposition and erosion including transverse aeolian ridges, depositional wind streaks, the north polar erg, laminated polar terrain as well as smaller erosional features like ventifacts, grooves, periodic bedrock ridges are observed by HRSC. Complex dunes fields are found to comprise barchans, barchanoid ridges, transverse dunes, reversing dunes, star dunes and domes indicate complex multi-directional wind regimes. Dark multiple layers in impact crater point to the source dark dune fields that might have been locally derived from eroded outcrops of these lithified dark-toned layered rocks. Over its operational period HRSC images have captured multiple instances of dust devils with diameters of 50 meters or greater. Comparison of the observed forward velocities with the wind regime predicted by the Martian Climate Database (MCD) show that the vortex directions were generally consistent with the modelled winds.