The eastern rim region of Hellas basin is characterized by the four prominent and quite extensively researched (cf. Crown et al., 2005, and references therein) outflow channels. In this work we focus on the Reull Vallis. On the basis of observations from available data sets, we present a hypothesis for the evolution of Reull Vallis and its complimentary fluvial system. We suggest that this system consists of parts that were formed during several phases rather than being a single continuous channel. Our results show that the fluvial system of Reull Vallis consists of two main parts and likely had independent formation phases and different sources of water. Our results also show that the upper portion of the Reull Vallis was formed by outflow from beneath Hesperia Planum (as proposed already in earlier works), but the suggested segments 1 and 2 (Mest and Crown, 2001) of the Vallis are not directly linked. There seems to have been an on-surface source for the formation of the segment 2 in the form of a topographic depression that was filled before the subsequent draining and formation of segment 2. Our interpretation of the evolution and formation implies a complex history for the Reull Vallis system.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 We defined and analyzed the different parts of the fluvial system and correlated temporally the processes that led to their formation using available images and topography data (Viking MDIMs, Mars Observer Camera (MOC), Thermal Emission Imaging System (THEMIS; both IR and VIS), the High-Resolution Stereo Color (HRSC) imager of Mars Express, and MOLA-gridded topography (128 pixels/degree)). Crater counting within the studied regions served to derive relative and model absolute ages of different portions of the Reull Vallis system, and measurements from MDIM and MOLA were used to estimate volumes and their balances for different parts of the system. THEMIS IR was used to get a comprehensive and detailed view of the different parts of the fluvial system; a survey of finer details was done from MOC NA and HRSC images. THEMIS multispectral data was also used in some of the channel areas, but analyzing their data is beyond the scope of this work.
2. Crater Statistics
 One of the first tasks of our work was to estimate the crater retention ages within different regions linked by the Reull Vallis fluvial system and compare relative ages of these regions and assess a broad time framework of the evolution of Reull Vallis. In order to accomplish this task we determined the number and size distribution of the superposed craters on the determined geological areas [Tanaka, 1986]. For our calculations we defined and measured three large areas (Figures 2a–2c) using observed geological and topographic boundaries, where we counted craters on the base of the Viking MDIM images (resolution is ∼256 pixels/degree). We named these study areas as Hesperia Planum area (Figure 2a), Morpheos basin area (Figure 2b), and Reull Vallis area (Figure 2c).
 1. The Hesperia Planum area, ∼1.5 × 106 km2, is completely within Hesperia Planum and corresponds to the central portion of this region covered by the Hesperian ridged plains (unit Hr [Greeley and Guest, 1987]). The ridged plains and the age derived from the count should provide the upper time limit for the formation of Reull Vallis because the uppermost portions of Reull Vallis cut the plains.
 2. The Morpheos basin area, ∼0.24 × 106 km2, is located in the southern portion of Hesperia Planum and corresponds to an extensive topographic depression in this portion of the Planum. The area could be defined by geologic contact of the smoothened terrain and the surrounding highland. We propose the provisional name “Morpheos basin” for this basin, and use it in this paper for clarification. The basin may have served as a place of accumulation either for effluents from segment 1 or smooth plains material from an unknown source.
 3. The Reull Vallis area, ∼0.27 × 106 km2, is a broad region on both sides of segments 2 and 3 that includes smooth plains around Reull Vallis. This area includes the majority of the unit Hps and part of AHh5 unit identified by Mest and Crown . Outcrops of the ancient cratered terrains were excluded from the area of crater counting.
 The crater counting resulted in total 3884 craters in the diameter range from 1.2 to 49.9 km. Craters less than 1.2 km were not included due to the limitations of the resolution and in order to minimize the possible inclusion of secondary craters. Clusters of secondary craters were also excluded from the counts. The size frequency distribution data is summarized in Figure 2d and Table 1; N(5) represents the cumulative numbers of craters with >5 km/106 km2, which were compared to the presented general stratigraphy of Mars [Tanaka, 1986]. The crater size-frequency distributions derived from the crater counting show that the curves for the Morpheos (Figure 2b) and Hesperia (Figure 2a) areas practically coincide with each other (Figure 2d) and correspond to the lower Hesperian epoch [Tanaka, 1986]. The curve for the designated Reull Vallis area (Figure 2c) is distinctly lower (Figure 2d) and corresponds to the transition from the lower to upper Hesperian epochs [Tanaka, 1986]. This is a result of the extensive resurfacing and later modification of this region. The formation of the Vallis itself predates this resurfacing. The counts are in good comparison with the results of the Mest and Crown  study, besides the differences in areal coverage and number of units. This is true for the Hesperia Planum and Reull Vallis area, but the geological mapping of Mest and Crown  shows only a very small portion of the Morpheos basin. In general, the derived counts of both studies are consistent with the proposed framework for the general evolution of the system presented in this study.
Table 1. Results of Crater Counting Within the Three Selected Areas Shown in Figures 2a–2c
Surface Area, km2
Number of Craters
Measured Crater Size Distribution, km
1.50 × 106
0.24 × 106
0.27 × 106
3. General Characteristics of the Reull Vallis Fluvial System
 The topographic profile along the thalwegs of the channels that constitute entire Reull Vallis system (Figure 3) shows that the system consists of two major parts. First, the lower and morphologically coherent portion of the system (segment 3; Mest and Crown ) extends for about 400 km and occurs in the elevation range from ∼−2,4 to ∼−3,2 km. Second, the more complex upper portion of the system is much longer (about 1300 km), spans the elevation range from ∼1.2 to ∼−2.4 km, and mostly corresponds to segments 1 and 2 of Reull Vallis [Mest and Crown, 2001]. The average topographic gradient ∼0.12–0.14° characterizes segment 2 and the major portion of the upper parts Reull Vallis. That, however, ends by a shorter (∼120 km) and steeper (∼0.77°) section, a kink-like feature (Figure 3), that connects the upper and the lower portions of the Reull Vallis system. Specific morphologic and topographic features of the upper portion of the system allow it further subdivision into smaller parts.
3.1. Northern Trough
 The topographically uppermost feature of the Reull Vallis system is a linear trough within Hesperia Planum immediately to the north of segment 1 (Figure 4a). Some details of this northern trough were noted by Mest and Crown , although the full extent could be observed only from MOLA data. This feature appears to belong to the Reull system because (1) it directly continues the topographic trend of segment 1 of Reull Vallis to the north (Figure 4b); and (2) it is a unique topographic feature atypical for this portion of Hesperia Planum (Figure 4b). The trough is a U-shaped depression tens of kilometers wide and 100–200 m deep. It begins at about 247°W, 29°S in a broad (∼60 km across) and shallow (∼100 m deep) depression and runs to the south for about 140 km where it joins segment 1. The trough becomes progressively narrower and deeper southward (∼40 km wide and ∼200 m deep at its southern end); the total volume of the feature is estimated to be about 430 km3 (Table 2).
Table 2. Volumes of Different Parts of the Reull Vallis Fluvial System
Part of Reull Vallis System
Morpheos Basin (up to 650 m contour)
 The trough has a little morphologic expression and represents almost a pure topographic feature, the surface of which is morphologically indistinguishable from the surrounding ridged plains (Figure 4a). Within the lower third of the trough, however, there is a narrow (2–3 km wide) and quite long (∼120 km) channel [Mest and Crown, 2001] that cuts the thalweg of the trough and breaches a small impact crater near the beginning of segment 1. A longitudinal topographic profile within surrounding plains parallel to the general orientation of the trough and segment 1 shows that the surface of the ridged plains near the trough is roughly horizontal (Figure 4c) while the thalweg of the trough displays a steady southward slope (∼0.10°), which is the same as the average slope along the thalweg of segment 1.
3.2. Segment 1
 Segment 1 of Reull Vallis starts as full-sized and steep-sided canyon at the southern end of the trough (Figure 4a). The source area of the segment occurs at the major break of slope of the surface of the ridged plains from the roughly horizontal to the steady southern slope (Figure 4c). The main channel of segment 1 begins as a relatively shallow (∼100–150 m) feature that deepens in its middle portion to ∼200 m and shallows again to ∼100–150 m near its end (Figure 5).
 Segment 1 clearly cuts the surface of the plains and displays all sorts of features consistent with the surface runoff (Figures 6a and 6b; see also Mest and Crown  and Crown et al. ). Rubbles of flat-topped angular to rounded blocks that represent pieces of the Hesperian lava plateau occur inside of the main canyon near its walls. A dense network of narrow fractures that extends into the plateau outside of the canyon separates the blocks. These features suggest collapse of the lava plateau owing to undermining by a subsurface flow. Inside the canyon, the blocks are more rounded to streamlined, terraced, and eroded. In places, there are bar-like features behind the blocks. The fragments of the collapsed plateau occur near the walls of the canyon where it broadens and are absent in its center where they were probably swept away by the flow through the channel (Figure 6a).
 The upper and middle portions of segment 1 are branching (Figure 6a). This means that the main channel of the segment was over-banked and smaller and shallower secondary channels were formed outside of the main channel. On the floor of the main and secondary channels there are scour marks that often occur behind obstacles such as remnants of wrinkle ridges. A low scarp resembling a cataract terminates at least one of the secondary channels. Numerous streamlined flat-topped islands typically occur in the upper and middle portions of segment 1. The walls of the islands and the channels display terraces that are not paired and likely represent erosional features.
 The lower third of segment 1 is different from its upper parts both morphologically and topographically. Scour marks and streamlined islands disappear there and the channel becomes progressively shallower, wider, and less distinctive. At its southern end, segment 1 cuts the northern rim of a large impact crater and disappears as a morphologic feature inside the crater (Figure 6b). There is no evidence for the second breach of the rim that would indicate the outlet of segment 1. The total volume of material removed from segment 1 is estimated to be about 514 km3 (Table 2).
3.3. Morpheos Basin
 Segments 1 and 2 of Reull Vallis are not connected morphologically [i.e., Mest and Crown, 2001]. The area of the gap between the segments represents a compact closed flat-floored topographic depression, the “Morpheos basin” [Kostama et al., 2004; Ivanov et al., 2005] consistently outlined by 650- to 700-m contour lines (Figure 7a). The deepest portions of the basin floor are at about 450 m, which is about 50 m deeper than the level of the floor at the beginning of segment 2. Volume of the Morpheos basin within the contour 650 m is estimated to be ∼17,000 km3 (Table 2).
 The morphologically smooth floor of the basin is surrounded from the south, west, and northwest by the rugged surface of the cratered terrains. The contact between the plains on the floor and the cratered terrains is sharp and preferentially occurs along the 650-m contour line across the basin (Figures 7b–7d). The same level marks the terminus of segment 1 (Figure 7b), the ends of two small channels entering the Morpheos basin from the northwest (Figure 8a), and the transition from the rugged to smooth morphology in the westernmost portion of the Morpheos basin where segment 2 of Reull Vallis begins. Wrinkle ridges on the floor of the basin appear to be morphologically subdued and less abundant than within the ridged plains on both sides of segment1 and elsewhere in Hesperia Planum [Ivanov et al., 2005]. Lobate ejecta that suggest the presence of a volatile-rich target characterize some of impact craters on the floor of the basin. These morphologic and topographic characteristics of the Morpheos basin are consistent with emplacement of sedimentary materials on its floor and formation of horizontal and morphologically smooth plains there [Mest and Crown, 2001].
3.4. Segment 2
 Segment 2 constitutes the midportion of the Reull Vallis fluvial system and represents a steep-sided canyon about 10–20 km wide and 200–400 m deep on average. Segment 2 appears as a faint morphologic feature within the Morpheos basin near its western edge and suddenly becomes a full-sized channel ∼10 km wide and 250 m deep outside of the basin (Figure 8a). Breaching the edge of the basin, the channel turns to the south and cuts through a large impact crater leaving both the inlet and outlet of the channel well preserved at the northern and southern rim of the crater (Figure 8b). There are implications of deposits within the crater that are cut by the segment 2. This could imply that the formation of the segment 2 may have occurred in several episodes (e.g., filling of the crater prior to the breaching of the southern wall) or that the segment 2 has been a channel for prolonged and variable fluvial activity.
 In contrast to segment 1, segment 2 lacks such features as streamlined islands and secondary branching channels that are suggestive for a catastrophic outflow. In the uppermost portions of segment 2, however, there are erosional features such as nonpaired terraces and scour marks on the floor (Figure 8b). These features disappear soon as the channel leaves the crater and turns to the SW at about 39.6°S, 111.2°E. After this point, the floor of segment 2 is completely covered by lineated material and the scour marks disappear but some terraces on the walls are still observable. Along the major portion of the segment, however, its walls appear to be sharp, smooth, and, in places, scalloped. Thus the morphology of segment 2 is clearly different from that of segment 1 and many features suggestive for the catastrophic flow are either absent or restricted to the upper portions of segment 2.
 The depth of the channel of segment 2 (defined as the difference between the median elevation of the surface on both banks and the deepest point on the floor) becomes progressively larger along the course of the segment (Figure 9). There are three distinctive regions on the plot. Within the upper portion of the channel, its depth steadily increases from ∼100 m at the Morpheos basin to ∼300 m at the point where the channel turns to the southwest. In the middle portion of segment 2 the depth of its channel displays significant variations but appears to oscillate around 300 m. The lower portion of segment 2 is characterized by rapid and significant increase of the channel depth from ∼400 m (at about 42.6°S, 104.4°E) to about 950 m in the area where segment 2 merges with segment 3 (Figure 9).
 The topographic profile along the thalweg of segment 2 displays significant variations with amplitudes up to several hundred meters that are probably related to different amount of material that fill the channel. The main topographic excursion is along the thalweg of segment 2, the kink-like feature (Figures 3 and 9), cannot be explained by the late filling. The kink characterizes the lower portion of the segment and represents the steepest (average slope is ∼0.77° over the distance of ∼120 km) section of the floor within entire Reull Vallis. Before and after the kink, the general topographic gradient along the thalweg of the channel is ∼0.11°–0.14° over the distance of many hundreds of kilometers. The kink corresponds to the region of the rapid increase of the channel depth and begins in the area where broad banks of segment 2 are narrowing and major topographic ridges of the cratered terrains come close to the channel (Figures 10 and 11a). The kink portion of the channel cuts through these ridges of the cratered terrain, and the channel gets wider and deeper within the area of the kink-like feature (Figure 11b). It is important to note that the kink occurs exclusively within the channel of Reull Vallis and a major topographic step that may correspond to the kink and control its formation does not occur in a broad region around Reull Vallis. The regional topography on both sides of Reull displays a steady slope toward the Hellas basin without large and consistent breaks of slope (Figure 12). The only break in slope in the southern bank is caused by the formation of the tributary to the segment 2 (Figure 11c).
 The volume of segment 2 is estimated to be ∼1,700 km3 (Table 2). This value represents the minimum volume of the channel because it is extensively covered by late debris flows [e.g., Ivanov et al., 2005].
3.5. Segment 3 and Teviot Vallis
 Segment 3 is the lowest and most impressive portion of the Reull Vallis fluvial system. It begins at 42.7°S, 102.5°E in the area where segment 2 and Teviot Vallis merge together (Figure 11a). The topographic profile along Reull Vallis (Figure 12) shows that segment 3 begins after the lower end of the kink-like feature characterizing the lowermost stretches of segment 2 (Figure 13). The general slope along the thalweg of segment 3 is much shallower than the slope within the kink-like feature and is similar to the slope characterizing the upper portions of the Reull Vallis. Shorter (∼140 km long) and longitudinally oriented Teviot Vallis is usually considered as a tributary or a “side canyon” [Mest and Crown, 2001] to the much longer and latitudinally oriented Reull Vallis. The topographic and morphometric configuration of segments 2, 3, and Teviot Vallis suggests, however, that segment 3 and Teviot Vallis form a single system of very deep canyons and segment 2 joins this system (Figure 12).
 Segment 3 differs from the other parts of the Reull Vallis fluvial system in many aspects. Its channel is significantly wider and deeper, especially in the beginning, and more sinuous comparing with the upper segments of Reull Vallis [Mest and Crown, 2001; Crown et al., 2005]. Streamlined islands, terraces on the walls, and secondary branches are absent, and edges of the channel are etched and gullied. In its beginning, segment 3 represents a very deep (∼1.5 km) and broad (∼50 km) canyon, which becomes more narrow and shallow downstream but still is deeper and wider than the other segments of Reull Vallis. The morphometric parameters of the upper portion of segment 3 (width, depth, and cross section) are drastically different from those characterizing the adjacent segment 2 but very similar to the parameters of Teviot Vallis (Figures 13 and 14), the floor of which continues the topographic trend of the floor of segment 3 (Figure 12). The floor of segment 3 is mostly covered by debris flows that flowed into the channel (Figures 11a and 13) preferentially from its banks. The actual volume of the segment is estimated to be ∼8,300 km3, which, owing to late filling, is the lower bound for the volume of material removed from the channel (Table 2).
 The depth of segment 3 becomes progressively smaller from ∼1.8 km at the beginning of the segment to ∼0.8 km at its end (Figure 15). This is opposite to the trend that characterizes segment 2 (Figure 9) but similar to the general trend of segment 1 (Figure 5). The late filling of segment 3 by debris flows clearly changed the depth of the channel. Near the end of segment 3 and in the area around its terminus, a material with low albedo is exposed. The darker material has different morphology and clearly predates the debris flows and apparently represents a deposit on the floor of segment 3, which is not covered by the debris aprons. The depth of the channel in this area is, however, ∼0.8 km, which is about 1 kilometer smaller than the depth in the beginning of segment 3 where the floor is completely covered by debris flows. Dispersed viscous flows (debris aprons, flow-like deposits) [Squyres, 1979; Crown et al., 1992; Pierce and Crown, 2003] probably manifest the final hydrologic events and thus, the late filling of the channel cannot explain the long-wavelength trend of the depth changes and probably was responsible only for the local variations of the channel depth.
4. Discussion: Sequence of Events During Formation of Reull Vallis
4.1. Northern Trough
 The uppermost part of Reull Vallis consists of two distinctly different features, the northern trough and segment 1. Because the trough is a unique topographic feature in the southern portion of Hesperia Planum and continues the trend of segment 1 (both geographically and topographically; Figure 4) we interpret this feature as the uppermost part of the Reull Vallis fluvial system with its source beneath the Hesperia Planum. Mest and Crown  did note the presence of theater headed canyons, which are located within the trough that converged in a large basin at the beginning of the segment 1. The U-shaped and flat-floored trough-like structures characterize the uppermost portions of Dao and Niger Valles and have been interpreted as manifestations of a subsurface flow of fluidized regolith that removed material supporting the composite layer of the Hesperian ridged plains along the flow path and led to the subsidence of the surface of the plains [Baker, 1982; Squyres et al., 1987; Crown et al., 1992]. The similarity of the shape of the trough and its position relative to the open channel of segment 1 to the characteristics of the trough-like features at Dao and Niger Valles suggest that the same mechanism of a subsurface flow was responsible for formation of the trough.
 The topographic position of the trough at relatively high elevation implies that the subsurface flow through it represented the initial phase of formation of the Reull Vallis fluvial system (Figure 16). The complete absence of features that may indicate the surface runoff at higher topographic levels around the trough and the subsurface nature of the flow within the trough imply that the regolith below the ridged plains was the principal source of water.
4.2. Segment 1
 Segment 1 of Reull Vallis starts as a steep-sided deep canyon that clearly cuts the ridged plains (Figure 4a), and manifests the change from the subsurface to the on-surface flow. The beginning of segment 1 coincides with the major break of slope in the southern portion of Hesperia Planum (Figure 4c). These topographic and morphologic characteristics of segment 1 suggest that the regional topographic configuration was the important factor in the change of the nature of flow in the upper parts of the Reull Vallis system. Fragments of collapsed plateau suggesting the subsurface flow occur near the walls inside the canyon of segment 1. Thus the subsurface flow played some role in formation of the segment. The major features of the segment such as branching, streamlined islands and possible cataracts, and erosional terraces on walls strongly suggest that the main phase of segment 1 formation was related to a catastrophic outflow. This flow clearly postdates the subsurface flow because the fragments of the collapsed plateau are rounded, terraced, and, in places, streamlined. Thus the catastrophic outflow through segment 1 likely represents the next stage of formation of Reull Vallis.
 In its lower third, segment 1 is shallower, wider, lacks streamlined islands and branches, and display broad terraces on its flanks (Figures 6b and 7b). These characteristics suggest diminishing of the eroding power of the flow and its predominantly lateral incision and formation of broader and shallower channel instead of narrower and deeper canyon. Such a diminishing of the power could be due to flattening of regional slope but the topographic gradient is not changing radically both along the thalweg of segment 1 (including floor of the crater where it enters) and within the plains outside of it (Figure 4c).
 The large morphologic gap between segments 1 and 2 was explained as the result of coverage of the Reull channel by late plains material [Mest and Crown, 2001]. There are at least two major uncertainties with this explanation. First, the obvious sources of the covering material are absent; second, there is no evidence that would support the hypothesis that the channel is buried in the low areas, as some remnants should then be visible in the high-standing topographic details such as rims of impact craters. For example, segment 2 near its beginning crosses a large impact crater (Figure 8b) and the crater rim is breached on both sides, and the inlet and outlet of the channel are clearly seen. Segment 1 also enters a crater and breaches its northern rim (Figure 6b). In contrast to segment 2, however, the outlet of segment 1 is absent and this suggests that the crater represents a true terminus of segment 1 and its channel did not continue further.
 Several important features characterize the area of the gap between segments 1 and 2.
 2. The basin is outlined by the 650- 700-m contour lines (Figure 7a). The sharp change of morphology from the rugged surface of the cratered terrain to the smooth surface of the basin floor consistently occurs near the 650-m contour line (Figures 7b–7d).
 3. The termini of segment 1 and smaller channels entering the basin from the northwest (Figure 8a) also occur at the elevation 650 m (Figure 7b).
 4. The evidence for a continuous channel connecting segments 1 and 2 is absent in the high topographic features within the basin (Figure 6b).
 5. The distinct morphological characteristics of segments 1 and 2 imply discontinuity between the two segments.
 6. The craters with fluidized ejecta are more abundant on the floor of the basin than outside of it.
 7. Wrinkle ridges are less widespread and more subdued on the floor of the basin than within the ridged plains of Hesperia Planum.
 These features are consistent and collectively suggest a hypothesis that a transient reservoir of water existed in the western portion of the Morpheos basin. The reservoir was probably filled by the effluents of segment 1 and served as the source of water that later carved the segment 2. The filling of the Morpheos basin, thus, represents the next stage of the evolution of this part of the Reull Vallis fluvial system. The crater retention age on the Morpheos basin floor is indistinguishable from the age of the Hesperian ridged plains (Figure 2d). Thus the first stages of the evolution of the Reull Vallis fluvial system occurred during the early Hesperian epoch (Figure 16), postdating at least the southern part of the Hesperia Planum formation.
 The evidence for a catastrophic flow through segment 1 allows us to make a very crude estimate of a time interval, which is required to fill the Morpheos basin. Depending on the velocity of the flow, the duration of the filling of the basin is estimated to be 2.8–10.4 days (Table 3). Used rates (50–80 km/h) are comparable to the terrestrial catastrophic water discharge rates of the Channeled Scabland and the Chuja Basin floods [Baker, 1973; Baker et al., 1993], but smaller compared to the proposed peaks of the Martian flood rates of 109 m3 s−1 [Carr, 1996]. The lower velocities of the flow appear to be less likely because, owing to the lower gravity on Mars, the eroding power of the stream would be significantly smaller than on Earth. The residence time of the body of water in the Morpheos basin is unknown but it probably was short enough to preserve water mostly as liquid. However, we cannot say for sure whether it was only a single event that formed the upper parts of the Reull system, or were there several consecutive events. A fast flooding of the Morpheos Basin and formation of the segment 2 could give some indication to why there is not much age difference between Hesperia and Morpheos.
Table 3. Mean Time Intervals (Days) Required for Filling the Morpheos Basin to Different Contour Levels
Contour Line, m
Velocity of Stream, km/h
4.4. Segments 2 and 3 and Teviot Vallis
 Segment 2 runs mostly on a steady, shallow slope of ∼0.13° except for the short (∼120 km) section at the end where the slope rapidly increases up to ∼0.77° (Figures 10 and 11a). The width and especially the depth of the channel at this section of segment 2 also rapidly increase (Figure 11b). The kink-like feature appears to play the key role in the understanding of evolution of the Reull Vallis fluvial system.
 The surface of plains on both sides of segment 2 displays a steady southwestern slope without a major step-like topographic feature in the area of the kink (Figures 3, 9, and 12). Thus formation of the kink was not controlled by regional topography and likely represents a feature related to the changes of the flow regime through segment 2. The instant increase of the slope, width, and depth of the channel along the kink means the flow was faster and more powerful there. Without the regional topographic control, such a change of the character of the flow could be caused only by the existence of a new, lower level of erosion.
 Segments 2 and 3 are drastically different from each other by their morphometric parameters (Figure 15). In contrast, the upper portion of segment 3 and Teviot Vallis are very similar by these characteristics (Figure 15). The overall morphology (Figures 11a and 13), the distribution of major topographic features in the area where Teviot Vallis and segments 2 and 3 are merging (Figure 14), and the topographic trends along the thalwegs of these channels (Figure 12) strongly suggest that Teviot Vallis and segment 3 of Reull Vallis represent a single large-scale structure and that segment 2 is a tributary to it. Both Teviot Vallis and the beginning of segment 3 are at significantly lower elevation (about −2.4 km and lower) than the major portion of segment 2 before the kink (about −0.9 km and higher). Thus Teviot Vallis and segment 3 may have provided the lower level of erosion, which is necessary for formation of the kink. In this case, Teviot Vallis and segment 3 must predate both the flow through segment 2 and formation of the upper portions of the Reull Vallis fluvial system (Figure 16). This assumption is supported by the similar morphometric characteristics of the Teviot Vallis segment 3 (Figures 15a–15c), and the filling near the apparent terminus of the segment 3, which can be understood as a result of the diminishing supporting power of the flow from the segment 2 flowing through the preexisting segment 3 and the resulting sedimentation.
 The region around the kink differs from the other portions of segment 2. In this area, a large topographic ridge of ancient cratered terrain comes close to the channel of the segment from the north (Figure 10). On the southern bank, in the area corresponding to the kink there are outcrops of ancient cratered terrain that are heavily eroded and channeled (Figure 10). Several small-scale and sourceless channels cut the cratered terrain, run toward segment 2, and join it near the lower portion of the kink. The surface of smooth plains on the southern bank of the upper portion of segment 2 slightly rises toward the remnants of the cratered terrain to the area where the kink begins (Figure 11c). These morphologic and topographic characteristics of the region of the kink suggest that a topographic barrier of the cratered terrain existed on the way of segment 2. The ridge is well preserved on the northern bank but heavily eroded and almost completely erased on the southern bank of segment 2. The position of the ridge on the banks corresponds to the section of the kink inside the channel. Thus the ridge likely served as a dam against which the flow from the upper portion of segment 2 was pond. The dam was finally overfilled and the flow broke through it and rushed down to the lower level provided by the Teviot Vallis segment 3 system leaving behind a short, steep, and deeply incised channel within the kink section of segment 2.
 The crater retention age of smooth plains on both sides of segment 2 is distinctly smaller than within Hesperia Planum and in the Morpheos basin (Table 1 and Figure 2d) and corresponds to the boundary between the lower and upper Hesperian epochs [Tanaka et al., 1992; Hartmann and Neukum, 2001]. It is likely, however, that segment 2 formed soon after the filling of the Morpheos basin and the residence time of water in the basin was significantly shorter than several hundreds of million years. In this case, the younger age of the smooth plains should be due to later episode(s) of resurfacing. This episode may be related to water released from the Morpheos basin. The volume of the basin is about an order of magnitude larger than the actual volume of the channel of segment 2 (Table 2). The channel, however, is filled by debris material that in many cases flow into the channel from its banks. Even if we assume that the original channel was two times deeper (which appears to be an overestimate because the depth of the channel is not changed significantly between its empty and filled sections), the volume of material removed from it would be five times smaller than the volume of the Morpheos basin. The basin, most likely, did not contain pure water. Assuming that sedimentary load in water was about 40% by volume, as proposed by Komar  for the Martian floods (see also Carr ), the volume of pure water in the Morpheos basin would be three times larger than the doubled actual volume of segment 2. This excess of water in the apparent source of the segment may have overbanked its channel, flow on the banks, and saturated the regolith there. The water-saturated regolith could be later mobilized and flowed to partly fill the channel. The later flow of the regolith, thus, may have contributed to the resurfacing of plains on the sides of segment 2.
 It is also possible that the basin may not have been filled to its full extent before breaching the crater wall (Figure 8b). In this case, the net volume of water passing through the segment 2 channel would have been slower and more gradual after a small initial pulse when the crater wall was breached, reaching some steady state between the water draining out of the basin through the gap in the crater wall and the inflow of water via segment 1. As this slow draining occurred, the focusing of sediment-rich water into the crater wall gap would have enhanced downcutting, leading to the backcutting into the basin a short distance that is observed (Figure 8a). In this scenario, the net volume of segment 2 being smaller than the volume of the source region would not mean anything. There are still several places in the length of segment 2, where more local overflow may have occurred, and this excess water may have contributed to the saturation of the regolith.
 The analysis of morphology and topographic characteristics of the Reull Vallis fluvial system suggests that it consists of two principal parts. The first part of the system, the lower Reull, includes Teviot Vallis and segment 3, and represents a single wide and deep channel that lies at lower elevations. The second part, the upper Reull, includes the northern trough, segment 1, the Morpheos basin, and segment 2 connected by successive episodes of water release and storage. The catastrophic character of flow through segment 1, filling and discharge of the Morpheos basin, and the tributary character of segment 2 suggest that the upper Reull as a whole postdates formation of the lower portion of the Reull Vallis fluvial system (Figure 16).
5. Conclusions and Future Work
5.1. Configuration of the Reull System
 The fluvial system of Reull Vallis consists in fact of two main parts (cf. Figure 1a): the lower Reull that includes Teviot Vallis and segment 3 of Reull Vallis and the upper Reull that consists of the northern trough, segment 1, the Morpheos basin, and segment 2. Both parts of the entire Reull system are different by the topography, morphology, and morphometry. The parts also likely have independent histories and different sources of water, although the source for the lower Reull remains unknown. To the contrary of the former interpretations, the upper Reull appears to be a tributary (Figures 12–14) to the lower Reull, the older part that existed before formation of the upper Reull.
 There are two important points made for the sources of the different parts of the system: First, materials below the composite layer of the ridged plains that make up the surface of Hesperia Planum appear to be the source of water that carved the upper portion of the Reull Vallis fluvial system and supplied the next part (Morpheos basin) with water. Second, there is also good evidence to support the suggestion that the segment 1 and 2 are not directly linked, but in fact, there was an on-surface source for the formation of the segment 2, the Morpheos basin, which was filled and later drained through the highlands and effectively carving segment 2. The ongoing Mars Reconnaissance Orbiter and Mars Express mission could very well provide the needed reassurance with their high-resolution instruments; HiRISE and HRSC for this, as well as details of the other parts of the system (e.g., lower and upper Reull connection or the connection with Harmakhis Vallis).
 From the measured properties of the basin and segment 1, some assumptions for the filling of the Morpheos basin may be made; if a single event was responsible, the filling of the basin by a flow through segment 1 did not require much time.
5.3. Evolution of Upper Reull
 The history of the upper Reull consists therefore of five distinct episodes (Figure 16): (1) a subsurface flow within the northern trough (Figure 4a); (2) a catastrophic outflow that carved segment 1 (Figure 6a); (3) filling of the Morpheos basin (Figure 7a) by the flow through segment 1; (4) discharge of the Morpheos basin and formation of segment 2 (Figures 8a and 8b); and (5) formation of the kink-like topographic feature connecting the upper Reull system to the lower Reull system.
5.4. Age Determinations
 The northern trough and segment 1 deform and cut the surface of Hesperian ridged plains (unit Hr) and empty into the Morpheos basin. The crater retention age of the basin floor, however, is indistinguishable from the age of the ridged plains (Figure 2d). The residence time of water within the basin was likely short in order to preserve water as a liquid that carved segment 2. Thus the formation of the upper Reull as a whole occurred in the very beginning of Hesperian epoch.
 The younger (late Hesperian-early Amazonian) age of the smooth plains on the banks of segment 2 could be explained by late, postchannel, episode(s) of resurfacing. The resurfacing, at least partly, may have been related to the viscous flow of water-saturated regolith within the smooth plains. The possible excess water from the Morpheos basin may have overfilled the channel of segment 2, and contributed to the saturation on both sides of the segment.
 This study has been funded by the University of Oulu; the Academy of Finland (V.-P. Kostama and M. A. Ivanov); the Magnus Ehrnrooth Foundation (V.-P. Kostama); Vilho, Yrjö, and Kalle Väisälä Donation Fund of Finnish Academy of Science and Letters (V.-P. Kostama); and Russian Fund for Basic Research, grant 07-05-00495-A (M. A. Ivanov). The Nordic Regional Planetary Image Facility (NRPIF) is acknowledged for the working facilities and Niko Koskensalmi for his participation in this work. The very useful reviews by Les Bleamaster and Vinny Gulick are much appreciated and improved our work significantly. We thank the High-Resolution Stereo Color (HRSC) Experiment Teams at Deutsches Zentrum fur Luft- und Raumfahrt Berlin and Freie Universität Berlin; the Mars Express Project Teams at Estec; and European Space Operations Centre for their successful planning and acquisition of data, as well as making the processed data available to the HRSC team. We acknowledge the effort of the HRSC Coinvestigator Team members and their associates, who have contributed to this investigation in the preparatory phase and in scientific discussions within the team. We also acknowledge the efforts by the Mars Orbiter Laser Altimeter (MOLA), Mars Orbiter Camera (MOC), and Thermal Emission and Imaging Spectrometer (THEMIS) science teams for making their data available to the public.