Drainage reversals, an end-member case of drainage reorganization, often occur toward cliffs. Reversals are commonly identified by the presence of barbed tributaries, with a junction angle >90°, that preserve the antecedent drainage geometry. The processes that form reversed drainages are largely unknown. Particularly, barbed tributaries cannot form through a spatially uniform migration of the cliff and drainage divide, which would be expected to erase the antecedent drainage pattern, and tectonic tilting toward the cliff that could reverse the flow direction is inconsistent with geodynamic models of large-scale escarpment, where many reversals are documented. Here, we propose a new mechanism for drainage reversal, where the slope imbalance across a cliff, together with the high erodibility of sediments that fill cliff-truncated valleys, result in faster divide migration along valleys compared to interfluves. We demonstrate this mechanism along channels that drain toward the escarpment of the Arava Valley in Israel. Reversal is established by observations of barbed tributaries and opposite-grading terraces. We show that drainage reversal occurs when erodible valley fill exists, and that the reversal extent correlates with the thickness of this fill, in agreement with the predictions of the proposed mechanism. This new reversal mechanism demonstrates that valley fill could play an acute role in fluvial reorganization processes, and that reversals could occur independently of tectonic tilting.

Abstract. The width of valleys and channels affects the hydrology, ecology, and geomorphic functionality of drainage networks. Valley and channel widths are often estimated through a power-law scaling between width (W) and drainage area (A), and where lithologic variability or differential uplift rates dominate, width was suggested to scale with both slope (S) and drainage area, through the relation W = kb Ab Sc. However, in fluvial systems that experience drainage reorganization, abrupt changes in drainage area distribution can result in widths that are disproportional to their drainage areas. Consequently, in such cases, the width-area-slope scaling is expected to deviate relative to drainages that did not experience reorganization. To explore the effect of reorganization on width-area-slope scaling, we studied 12 valley sections in the Negev desert, Israel, categorized into undisturbed, beheaded, and reversed valleys. We found that the drainage area exponent, b, differs between valley categories, and that reversed valleys are characterized by a negative b exponent, indicating valley narrowing with increasing drainage area. A detailed study of a reversed valley reveals that unlike the negative b exponent that links drainage area to valley width, the relation between drainage area and channel width is best fitted with a positive b exponent. This difference indicates that the timescale of channel width adjustment to post-reorganization drainage area distribution is faster than that of the valley width adjustment. We find that the difference in channel width across the divide causes a step change in unit stream power between the adjusted reserved channel and the unadjusted, beheaded channel. Gradients in width and unit stream power across the divide, lead to a width-feedback that promotes ongoing divide migration and reorganization. The identified distinct width-area-slope scaling of reorganized valleys could assist in recognizing and constraining the dynamics of landscapes influenced by drainage reorganization and likely has critical implications for the distribution of erosion rates in reorganized landscapes.

<p>Observations from around the globe show that drainage reversal toward cliffs (and at a larger scale, toward escarpments) is a common phenomenon.  Drainage reversal occurs when a channel that used to grade in one direction reverses its gradient while exploiting its antecedent valley, forming barbed tributaries with junction angle >90°. Drainage reversal is an important end-member of fluvial reorganization that drastically shifts the hydrologic and geomorphic functionality of the landscape.  The processes that induce drainage reversals, however, remain largely enigmatic. In many cases, tectonic or structural tilt of the surface is invoked to explain reversal toward the tilt direction, but independent evidence for tilting is rare. Moreover, in great escarpments, geodynamic models predict tilting away from the escarpment, opposite to the sense of reversal discussed here.</p><p>We study drainage reversals toward the southern Arava Valley escarpment in Israel, along the Sinai-Arabia transtentional plate boundary. In this area, we establish reversals by observations of barbed tributaries, valley-confined windgaps, and terraces and interfluves that grade opposite to the grading direction of the active channel. Detailed morphological and geological analysis of the field area gives rise to a new, tilting independent mechanism for drainage reversal toward cliffs. The initial condition for this mechanism is a cliff that truncates fluvial channels that flow over the highland and away from the cliff, and a water divide that coincides with the cliff. The truncated channels appear as saddles along the cliff and are commonly filled with alluvial and colluvial sediments. Such initial conditions characterize shoulder-type great escarpments and cliffs that form following river capture events. Importantly, in these settings, the sediments that fill the truncated channels are more erodible than the bedrock that builds the interfluves.</p><p>According to the mechanism we propose, the erodible valley fill near the steep cliff is initially transported down the cliff via hillslope processes, which results in a gradual migration of the divide along the antecedent valley and away from the cliff. A reversed channel segment forms between the receding divide and the cliff, such that along the channel, the divide and the cliff are not coincident anymore. The faster fluvial incision in the reversed segment with respect to the antecedent channel further pushes the divide away from the cliff. When the receding divide traverses a tributary confluence, a barbed tributary forms. The increased discharge of the reversed segment facilitates cliff embayment that eventually affects cliff retreat and morphology.</p><p>This new mechanism indicates that a relatively thin layer of erodible valley fill could be a tipping point that completely changes the trajectory of landscape evolution via drainage reversal. Importantly, however, flow reversal towards cliffs does not necessitate such a layer but instead could be triggered by other hydrological and geological conditions that promote faster erosion toward the cliff within the antecedent channel with respect to the interfluves. </p>

<p>Drainage reversals occur when a channel reverses its flow direction by 180<sup>o</sup> while exploiting its antecedent valley. This reorganization mode can critically impact landscapes' hydrologic and geomorphic functionality, but the processes inducing reversals and the related landscape dynamics have not been studied in detail. Reversals are commonly attributed to tectonic tilting. However, in many cases, independent evidence for tilting is missing. Furthermore, when reversals occur toward great escarpments, as was documented in many terrains around the globe, isostatic tilting is expected to occur away and not toward the escarpments.</p><p>The current study explores a natural laboratory for drainage reversals in the southeastern Negev Desert, Israel. We identified in this field area tens of highland channels that reversed their flow direction eastward and toward the Arava Valley Escarpment.</p><p>Reversals are established by observations of (1) barbed tributaries that join the main trunk with junction angles > 90<sup>o</sup> and preserve the antecedent pre-reversal drainage topology; (2) a valley confined drainage divide (windgap) that separates the reversed channel from the antecedent, beheaded channel; and (3) series of terraces that grade west, toward the windgap and opposite to the active flow direction.</p><p>Based on field observations and morphometric analysis, we propose a new, tilting-independent, mechanism for inducing flow reversals. According to this mechanism, reversal is linked to localized windgap migration within the antecedent channel and away from the escarpment. Migration is driven by slope imbalance across the windgap with steeper slopes at the escarpment side, and by erodibility differences between the hard rocky interfluves and the more erodible valley fill.</p><p>Using a new algorithm for quantifying valley width, we find that the scaling between drainage area (A) and width (W) differs between reversed, beheaded, and non-reorganized valleys. In addition to providing markers for reorganization, the unique A-W scaling leads to feedback that promotes further windgap migration and generates longer reversals.</p><p>The oppositely grading terraces that accompany some reversed channels present an outstanding opportunity for quantifying the dynamic and rate of windgap migration. We hypothesize that the abandonment age of each terrace reflects the timing at which the paleo-windgap migrated past the location of the terrace, promoting the incision of the reversed channel that generated the terrace. Accordingly, the abandonment ages of the terraces can inform us about the timing and dynamics of windgap migration.</p><p>Absolute ages of terrace abandonment were constrained by luminescence dating, with complementary relative ages inferred from chronosequence of reg soils, which develop on abandoned terraces in hyper-arid environments. In agreement with the reversal model, we found that the degree of soil development and the abandonment ages of terraces increase with distance from the windgap eastwards. The average windgap migration rate has been ~1 mm/yr since 200 Kyr, an order of magnitude greater than the vertical incision rate of the reversed channel. The age-distance relations of the terraces indicate episodic windgap migration with a recent stalling. A similar age for other dated windages in the region hints at a regional, possibly climatic control on windgap migration.  </p>

The copyright of individual parts of the supplement might differ from the article license. S1 Algorithm for identifying optimal locations for width measurementsThis section describes the semi-automatic ArcGIS-based algorithm for measuring valley width. The algorithm input is a valley bottom (VBET) polygon, and the output is a set of width measurement locations and values.

Abstract. The width of valleys and channels affects the hydrology, ecology, and geomorphic functionality of drainage networks. In many studies, the width of valleys and/or channels (W) is estimated as a power-law function of the drainage area (A), W=kcAd. However, in fluvial systems that experience drainage reorganization, abrupt changes in drainage area distribution can result in valley or channel widths that are disproportional to their drainage areas. Such disproportionality may be more distinguished in valleys than in channels due to a longer adjustment timescale for valleys. Therefore, the valley width–area scaling in reorganized drainages is expected to deviate from that of drainages that did not experience reorganization. To explore the effect of reorganization on valley width–drainage area scaling, we studied 12 valley sections in the Negev desert, Israel, categorized into undisturbed, beheaded, and reversed valleys. We found that the values of the drainage area exponents, d, are lower in the beheaded valleys relative to undisturbed valleys but remain positive. Reversed valleys, in contrast, are characterized by negative d exponents, indicating valley narrowing with increasing drainage area. In the reversed category, we also explored the independent effect of channel slope (S) through the equation W=kbAbSc, which yielded negative and overall similar values for b and c. A detailed study in one reversed valley section shows that the valley narrows downstream, whereas the channel widens, suggesting that, as hypothesized, the channel width adjusts faster to post-reorganization drainage area distribution. The adjusted narrow channel dictates the width of formative flows in the reversed valley, which contrasts with the meaningfully wider formative flows of the beheaded valley across the divide. This difference results in a step change in the unit stream power between the reversed and beheaded channels, potentially leading to a “width feedback” that promotes ongoing divide migration and reorganization. Our findings demonstrate that valley width–area scaling is a potential tool for identifying landscapes influenced by drainage reorganization. Accounting for reorganization-specific scaling can improve estimations of erosion rate distributions in reorganized landscapes.

No abstract

<p>Valley width is a fundamental morphologic property of rivers that plays a key role in drainage networks' hydrology, ecology, and geomorphology. In many cases, defining and measuring valley width is far from trivial. Therefore, similar to channel width, the valley width (<em>W</em>) is commonly approximated as a power law function of the drainage area (<em>A</em>) and expressed as <em>W = k<sub>c</sub>A<sup>d</sup></em>. Global observations have shown that the exponent  (<em>d)</em> can vary widely but is typically ~0.5. However, in fluvial systems that have undergone drainage reorganization, gradual or abrupt changes in drainage areas along the valley could produce widths that are disproportionate to their drainage areas. As a result, the valley width - drainage area relationship in reorganized systems is expected to differ from undisturbed drainages that have not undergone reorganization.</p> <p>To test this prediction, we studied 12 valleys in the Negev desert, Israel, and classified them into three categories, based on field evidence and remote sensing data: (i) undisturbed valleys, which are minimally affected by reorganization; (ii) beheaded valleys, whose headwaters were beheaded; and (iii) reversed valleys, which have reversed their flow direction by 180 degrees while exploiting their antecedent valleys. Using a new semi-automatic tool to measure valley width on high-resolution DEMs, we calibrated the best-fit power law for each valley to explore the relationships between drainage area and valley width for each valley category.</p> <p>Our results show that the valley width-drainage area scaling in reorganized valleys deviated significantly from those in undisturbed valleys in our field area and global observations. The drainage area exponents <em>(d</em>) were lower in beheaded valleys compared to undisturbed valleys but remained positive. In contrast, reversed valleys were characterized by negative <em>d</em> exponents, indicating valley width decrease with increasing drainage area. For the reversed category, we also explored the independent effect of channel slope (<em>S</em>), where the valley width is <em>W = k<sub>b</sub> A<sup>b</sup>S<sup>c</sup>,</em> which resulted in negative and overall similar values of <em>b</em> and <em>c</em>.</p> <p>In one reversed valley section, we compared the scaling of valley versus channel width as a function of drainage area. We found that in contrast to the downstream narrowing valley, the channel width shows an opposite trend and widens downstream, suggesting that the channel has mostly adjusted to the post-reorganization drainage area distribution. The narrow reversed channel shapes the width of the formative flows, which contrasts significantly with the wide flows of the beheaded valley across the divide. This difference results in a step-change in the unit stream power between the reversed and beheaded channels, potentially leading to a "width feedback" that promotes further divide migration.</p> <p>Our findings can be used to identify landscapes that have been affected by recent drainage reorganization and should be taken into consideration in studies that use the relationship between valley width and drainage area for valley width predictions, stream power calculations, and landscape evolution models.</p>