- Email: [email protected]
The resilience of a badland area to climate change in an arid environment
CATENA-01788; No of Pages 10 Catena xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
The resilience of a badland area to climate change in an arid environment Aaron Yair a, Rorke B. Bryan b, Hanoch Lavee c, Wolfgang Schwanghart d, Nikolaus J. Kuhn e,⁎ a
Hebrew University Jerusalem, Israel University of Toronto, Canada c Bar Ilan University, Tel Aviv, Israel d University of Potsdam, Germany e University of Basel, Switzerland b
a r t i c l e
i n f o
Article history: Received 17 February 2012 Received in revised form 5 April 2012 Accepted 15 April 2012 Available online xxxx Keywords: Badlands Model landscape Climate change Resilience
a b s t r a c t Badlands have long been considered as model landscapes due to their perceived close relationship between form and process. The often intense features of erosion have also attracted many geomorphologists because the associated high rates of erosion appeared to offer the opportunity for studying surface processes and the resulting forms. Recently, the perceived simplicity of badlands has been questioned because the expected relationships between driving forces for erosion and the resulting sediment yield could not be observed. Further, a high variability in erosion and sediment yield has been observed across scales. Finally, denudation based on currently observed erosion rates would have lead to the destruction of most badlands a long time ago. While the perceived simplicity of badlands has sparked a disproportional (compared to the land surface they cover) amount of research, our increasing amount of information has not necessarily increased our understanding of badlands in equal terms. Overall, badlands appear to be more complex than initially assumed. In this paper, we review 40 years of research in the Zin Valley Badlands in Israel to reconcile some of the conflicting results observed there and develop a perspective on the function of badlands as model landscapes. While the data collected in the Zin Valley clearly confirm that spatial and temporal patterns of geomorphic processes and their interaction with topography and surface properties have to be understood, we still conclude that the process of realizing complexity in the “simple” badlands has a model function both for our understanding as well as perspective on all landscape systems. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Badlands are often considered as an intensively dissected barren landscape, devoid of soil cover, with a sparse or absent vegetation (e.g. Howard, 1994). The very high drainage densities, V-shaped valleys and steep slopes, have often led to believe that badlands represent a landscape where the frequency and magnitude of overland flow and erosion processes are high to very high, resulting in rapid landscape evolution (see review by Wainwright and Brazier, 2011). This is why badlands have been sometimes regarded as “ideal field laboratories” for testing landscape evolution hypothesis (Bryan and Yair, 1982; Campbell, 1997; Schumm, 1956). Recent research has shed a critical light on the “simple landscape” perspective. Two key issues support this change in perspective. First, the erosion rates and sediment yields observed in badland areas vary widely. A recent comprehensive review, conducted by Nadal-Romero et al. (2011) of badland areas in Mediterranean environments, exhibits a huge variability in erosion rates. No relationship in erosion
⁎ Corresponding author. E-mail address: [email protected] (N.J. Kuhn).
rates was found for areas in the range of 0.0001–100 ha. What is more, erosion rates for areas as small as 0.1 m2 vary in the range of 1.1– 500 kg m− 2y− 1. A more pronounced variability (0.12–1000 t ha− 1y− 1) was obtained for larger areas covering 10–100 ha. For areas in the range of 100 ha to 10 km2 erosion rates are smaller, but still a ten-fold variability exists. The observed differences from hillslope to catchment scale are at least on the same order of magnitude than those typically observed in a wide range of drylands (reviewed in Branson et al., 1981). They are, apart from differences in lithology and topography, attributed to different patterns of sediment transport and deposition in the studied badlands and the scale effects of erosion and deposition (Nadal-Romero et al., 2011). No relationship was found between average annual rainfall and average erosion rates for annual rainfalls in the wide range of 80–1250 mm (Nadal-Romero et al., 2011). The limited effect of rainfall magnitude was also identified by Della et al. (2007), who stated that the threshold for runoff generation is 60 mm for 5 consecutive days. Below this threshold, surface runoff does not seem to reach the power required for significant modifications of the topographic surface. The observations in the studies cited above confirm data presented in comprehensive and detailed earlier papers (Della et al., 2007 and Gallart et al., 2002). Overall, the variability in sediment yields and limited relationship with annual rainfall indicate that the badlands are by no means simpler than more complex landscapes.
0341-8162/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.catena.2012.04.006
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
2
A. Yair et al. / Catena xxx (2012) xxx–xxx
A key issue limiting the use of badlands as simple, model landscapes appears to be the very high spatial variability in erosion rates observed under present day rainfall conditions. The spatial variability presents a serious barrier for the extrapolation of erosion rates from the hill slope to catchment scales and certainly to landscape evolution over longer geological time scales under changing climatic conditions. A second reason contradicting the simple geomorphic landscape is the mismatch observed between current erosion rates and the longterm denudation rates that have been reconstructed for several badland areas. For example, denudation rates observed by Yair et al. (1982) in the Zin Valley Badlands were on the order of 2–5 cm y − 1 and would be expected to completely wipe out an area with a relative elevation of 10 m within the very short time of 200 to 500 years. Similar mismatches have been reported for the Dinosaur badlands in Alberta (Hodges and Bryan, 1982) and to a lesser extent badlands in Tuscany (Torri et al., 1994). This is quite unreasonable unless we assume those denudation rates equal extremely high rates of tectonic uplift, or alternatively that high erosion rates result from extreme rain events with an extremely high frequency. A final reason for the presumed geomorphic simplicity of badlands may be their perception as a model landscape, which led to ignoring the complexity and spatial variability of surface processes in badland studies. Without singling out an individual study (for a selection of the methods used in badland research see Bryan and Yair (1982) and Torri et al. (2000)), we would just like to draw the attention to the way erosion was measured and extrapolated. Often runoff and sediment yield were monitored on plots and small catchments of varying size, often equating sediment yield at the lower end of the plot with erosion and thus ignoring deposition on the plot. This must invariably lead to high net erosion rates when plots are small compared to the average transport distance of the detached and entrained material (Kinnell, 2009). This problem is exacerbated because the selection of monitoring sites is often guided by the presence or absence of erosion features, without assessing their spatial representation for the entire badlands. In the light of this critical assessment of the idea of badlands as a “natural laboratory” for landscape development, 40 years of research activities in the Zin Valley badlands in the Negev desert are reviewed in this paper. The wide range of studies on different spatial and temporal scales illustrates the gradual development of our understanding of badlands and suggests some answers to some conflicting results mentioned above. The paper is split into two sections: first, the published work is reviewed, followed by some previously unpublished data acquired in the past five years focusing on linking processes to landscape development. 2. Study area The Zin Valley Badlands are located in the northern Negev desert (Fig. 1). They are incised into the clay rich shales of the Taqyia formation of Paleocene age. The field studies discussed in this review were conducted at the lower part of this formation where the clay fraction is 65–75% of the bedrock (Bentor, 1966). The clay mineralogy is dominated by 50% of Montmorillonite, 40% Kaolinite and 10% Illite (Arkin et al., 1972; Nathan, 1978). The present topography consists of generally V-shaped longitudinal valleys, with rounded ridges along the divides, straight sideslopes of approximately 30° angle, and rillto gully sized fluvial channels along the valley floor (Fig. 2). Average annual rainfall in the region, observed over 50 years, is 90 mm per year with extreme values of 34 and 170 mm, potential annual evaporation is 2500 mm (Yair and Kossovsky, 2002). Only few rainstorms yield more than 20 mm per day. Two well-defined terraces (Fig. 3) with an extensive scatter of prehistoric sites pointing at an intensive activity at the late Quaternary have been identified in the area (Yair et al., 1982). This is attributed to the extensive outcrop of the Mishash formation, which is rich in flint used for the preparation of working tools, and
Fig. 1. Location and average annual precipitation at the Zin Valley Badlands in the Negev desert, Israel.
the occurrence of nearby perennial springs along the Zin river. The lower terrace is situated below 400 m. Prehistoric sites here are attributed to the Upper Paleolithic period which span an interval of 45,000 to 25,000 years BP. The upper terrace (above 400 m) is composed of colluvial silts containing in situ Epi-Paleolithic sites dated by C 14 to 13,530 BP (Marks, 1977). The incision between the terraces ranges between 5 and 7 m, equalling average annual channel incision rates of 0.25–0.35 mm. These incision rates are an order of magnitude below those calculated based on erosion tests conducted in the past decades (see review of studies below) and illustrate that the conflicting results between short and long-term studies identified in the Introduction also apply to the Zin Valley Badlands. The prehistoric findings also clearly indicate that the main phase of erosion took place during the Upper Pleistocene when the level of the Dead Sea lake was approximately 200 m above its present level, representing a wet climatic phase, leading to the assumption that about 300 mm of annual rainfall were required to incise the badlands (Yair et al., 1980b). This adds a further complication to the interpretation of the data collected
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
3
N
Fig. 2. Aerial photograph of the western section of the Zin Valley Badlands. Note the longitudinal form of the badlands, indicating a relatively high rate of headward erosion. The size of the image is approximately 0.25 km2.
on current erosion rates in the Zin Valley Badlands: if erosion was higher during a phase of wet climate in the past, why haven't the badlands been eroded away completely a long time age if current erosion rates are still so high? 3. Surface characteristics, infiltration and erosion
Fig. 4. Crusts on north-(left) and south-facing slopes. Note the rough popcorn structure on the left and the smooth microtopography on the right.
3.1. Surface characteristics The slopes in the Zin Valley Badlands show roughly three distinct sets of surface characteristics (Figs. 4 and 5). On the ridges, a poorly developed regolith dominates. Above a layer of shards formed of physically weathered bedrock, a thin, 2 to 5 mm thick crust, broken by desiccation cracks, has formed. On north-facing sideslopes, a 20 to 30 cm deep, crusted soil with deep desiccation cracks and a “popcorn” surface microtopography with depressions of 4 to 8 cm depth
has developed (Yair et al., 1980a). On these slopes, seals of lichens are often found. Regolith on south-facing slopes is less developed than on the north-facing slopes, reaching a thickness of only 10 cm. Three distinctly different rill systems have developed on straight valley sideslopes and along concave valley heads (Figs. 4 and 6). On straight and plan north-facing slopes the surface is generally unrilled with the exception of desiccation cracks where small pipes can develop (Yair et al., 1982). On south-facing slopes, single, discontinuous rills dominate. Well-integrated rill systems are common along valley heads or concave slope segments where the topography of the slopes enhances the convergence of rills and the formation of rill networks. It is noteworthy that most rills begin near the ridges and do show only limited change in width and depth along the slope. 3.2. Small-scale erosion tests
Fig. 3. Topography and sites with Paleolithic artefacts on the upper and lower terrace of the Navarim contributory (dark line in upper left part of figure) to the Zin River (crossing S–N on right of figure).
Runoff and erosion on the different interrill and rill features have been studied using a range of infiltration and erosion tests. Two sets of sprinkling experiments were carried out. The first was conducted in 1979 (Yair et al., 1980a, 1982) and the second in 2001 (Kuhn and Yair, 2003; Kuhn et al., 2004). The first set was carried out on two different scales: small plots covering 1.5 m 2 and larger plots covering up to 45 m 2. The use of different plot sizes aimed at identifying potential scale effects on runoff and erosion. The sprinkling experiments on the small plots were conducted on opposite north- and south-facing slopes. Sprinkling on the small plots was conducted with the Morin rainfall simulation unit (Morin et al., 1970). This unit (Fig. 7) can reproduce accurately the intensity and kinetic energy characteristics of natural rainfall. Rainfall was applied at 35.3 mm h− 1 for 43 min, totalling 25.3 mm (Yair et al., 1980b). Runoff on the north-facing slope started after 5 min and increased gradually (Fig. 8). Total runoff yield was
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
4
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 5. Crust properties, regolith depth and Cs-137 activities along a Northwestern-facing slope section studied by Kuhn and Yair (2003).
11.98 mm representing a runoff coefficient of 47.4%. However, infiltration rate after 10 min was still quite high (27 mm h− 1), and 18 mm h− 1 after 20 min. Despite the high rain intensity no surface disturbance by splash was observed and no change in surface microtopography took place. Sediment yield was 667 g m − 2, representing a denudation rate of 0.48 mm, which must be considered as quite low for the high rain intensities and long durations applied. Rainfall at the south-facing plot was applied at 36.7 mm h− 1 and lasted 48 min, totalling 29.4 mm. Runoff started after 11 min (Fig. 8), increased gradually, and approached equilibrium conditions at minute 37. Runoff yield was 8.64 mm representing a runoff coefficient of
29.4%, much lower than that of the opposite slope. Infiltration rate, after 10 and 20 min, was still very high (37 mm h− 1 and 20 mm h− 1, respectively). Again, the change in surface characteristics during the 48-minute duration of the test was negligible. Sediment yield was low compared to the north-facing slope (38 g m − 2), representing a denudation rate of 0.027 mm. It is interesting to note that runoff and erosion on the north-facing slopes were greater than on those facing south, despite the much greater surface roughness and depression storage on the north-facing slopes. The greater infiltration on the south-facing slopes was attributed by Yair et al. (1980b) to the thinner, more porous crusts and shards underlying the thin top layer (Fig. 5). 3.3. Infiltration on crusts along a slope
Fig. 6. Rill systems and headward erosion on N-facing slopes in the Zin Valley Badlands. Straight and radial slopes show little or no rill development while on concave slopes integrated rill networks dominate. Undercutting by the main channel (lower right of image center) indicates that colluvial sediment is transferred into the main channel by mass wasting.
A second set of small-scale measurements were conducted on plots across slopes by Kuhn and Yair (2003) and Kuhn et al. (2004) using a needle sprinkler covering an area of 1 m 2 (Fig. 7). The tests aimed at bridging a gap in the previous observations by focusing on the effect of different crust characteristics observed on one slope on infiltration during quasi-natural rainfall event. The simulator produced drops of 3.5 mm diameter and a kinetic energy of 3.1 J m− 2 mm− 1 (Kutiel, 1978). These values are lower than for natural rainfall (Fox, 2004), but the existence and stability of a crust when rained on with the Morin unit described above indicates that this does not affect infiltration. Low intensity rainfall was applied for 15 min, followed by a high intensity rain burst (35 mm h− 1) for 10–12 min. This pattern simulates a natural rainstorm, when a high intensity rain shower occurs during a rainfall event with already wetted soil. The return interval of such an event is approximately 5 years in the Sede Boqer region (Yair and Kossovsky, 2002). Infiltration differed significantly for the different crust types (Fig. 9). Highest infiltration rates were obtained for the valley side slopes and the rills formed there (0.58 mm min− 1, not shown in Fig. 9 because only a final rate was observed). The ridges exhibited much lower infiltration and more
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
5
Fig. 7. The Morin (left) and Kuhn and Yair (right) rainfall simulator in the Zin Valley Badlands.
rapid runoff development than the side slopes (Fig. 9), indicating that partial runoff is generated during rainfall events of high frequency and low magnitude.
3.4. Slope-scale infiltration and erosion tests Slope‐scale infiltration tests were conducted using a sprinkler, which provided uniform coverage of 90 m 2 (Yair et al., 1982), but which does not accurately reproduce natural velocities of rainfall. The drop‐size spectrum is smaller than for natural rainfall and the fall height of the drops generates only 50% of the natural kinetic energy. However, like with the sprinkling tests conducted by Kuhn and Yair (2003), the existence of a stable crust indicated that the limited kinetic energy did not affect infiltration. The north-facing plot covered an area of 45 m 2, but runoff was collected only from the central 30 m 2. Average rain intensity applied over the sprinkled area was 48.5 mm h− 1. The test lasted 34.5 min, representing a rain amount of 28 mm. Runoff developed only in rills and displayed a pulsating character and varied in space and time (Fig. 10). Pipe flow was observed after 15 min and after 30 min rill bank and pipe collapse was widespread. The intermittency of the runoff is attributed to clogging of rills, either by the collapse of banks or deposition of sediment once rill infiltration losses reduce transport capacity of flow below the actual load. Similarly, collapsing pipes diverted runoff onto interrill areas where infiltration losses were high (Fig. 11). Runoff yield was 3.94 mm, representing a runoff coefficient of 14.2%, much lower than that obtained in the small plot. Due to the pulsating character of runoff, sediment concentration varied in the range of 32–172 g l− 1. Sediment yield was 235 g m− 2, representing an average denudation rate of 0.17 mm. All of the sediment originated from the rill banks and pipes along the rills. The south-facing plot covered also an area of 45 m 2, but runoff was collected from the central 30 m 2. Average rain intensity was only 18 mm h− 1. The test lasted 51.5 min (Fig. 10). Runoff started, on the ridge, after 5.5 min, and was quickly absorbed in desiccation cracks that did not seal. Continuous flow was limited to the rills, but did not occur simultaneously in all rills. The intermittent runoff generated only 9 l of discharge into the collecting tank, which corresponds to a runoff coefficient of only 1.7%. Sediment discharge was also very low (85 g m − 2), representing a denudation rate of 0.06 mm. 3.5. Surface characteristics and spatial patterns of runoff generation and continuity
Fig. 8. Runoff observed by Yair et al. (1980a) on small1.5 m2 plots with different aspect in the Zin Valley Badlands.
The infiltration tests show that runoff generation in the Zin Valley Badlands is highly variable in space and time. Runoff is generated more frequently on the ridges (30% of annual rainfall, than on the valley sideslopes (10% of annual rainfall) (Kuhn and Yair, 2003). This pattern leads to runoff discontinuity. A key question for understanding the complexity of badlands as a hydrologic and geomorphic system is the long-term interaction between runoff and infiltration patterns, the corresponding feedback on surface properties, in particular erosion
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
6
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 9. Infiltration rates observed by Kuhn and Yair (2003) during simulated rainfall on ridge and slope interrill areas in the Zin Valley Badlands. The error bars mark the maximum deviation from the mean observed during three replicate tests.
and deposition, and the consequences for landscape development. In the second part of this paper, an analysis of surface properties and topography aims at addressing the interaction between form and process and the long-term landscape development. The low infiltration on the ridges leads to rainflow erosion and limits soil development, maintaining soil susceptible to rapid sealing and thus creating a positive feedback on low infiltration (Kuhn and Yair, 2003). On the sideslopes, infiltration of rainfall and water flowing downslope from the ridges, cause flow discontinuity and deposition, which enhances soil development, leading to high infiltration and discontinuous runoff. The long-term effects of runoff generation and frequent discontinuous flow are reflected by soil development along the badland slopes. Soil depth is indicative of the balance between erosion and soil formation, being deeper in areas of low erosion, limited runoff and high infiltration, and shallow in areas with much runoff and erosion (Jenny, 1941). In the Zin Valley Badlands, weathering depth is
lowest at or close to the divide (Fig. 5). Downslope, weathering depth increases initially for 4 to 5 m, but declines again in the lower part of the slope. There is also a clear difference in weathering depth between non-rilled slope sections and those with an integrated rill system (Table 1). In the integrated rill system, weathering depth is generally shallower than in the single rill and almost constant along the entire
Fig. 10. Runoff observed by Yair et al. (1982) during slope-scale erosion tests in the Zin Valley Badlands.
Fig. 11. Discontinuous runoff, marked by disappearing dark dye tracer pointed out by white arrows, on north-facing slope during erosion test conducted by Yair et al. (1982).
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
slope, showing that, due to more frequent runoff, erosion rates are higher as well. Kuhn and Yair (2003) concluded that the frequent discontinuity of surface runoff is responsible for this pattern. Using the data from their infiltration tests, they estimated that rainfall intensities of as low as 10 mm per hour are sufficient to overcome infiltration losses in an integrated rill network while at least 25 mm are required for interrill areas. Dye tracers placed on ridges, interrill areas and in rills to monitor runoff during natural rainstorms support the results of the sprinkling tests when rainfall of 26 mm on December 6th 2001 was sufficient to produce continuous flow from ridges to valleys. However, even during this storm no surface runoff from the interrill areas on the sideslopes reached the rills (Kuhn and Yair, 2003). 4. Effect of partial area runoff generation and discontinuity on landscape development 4.1. Cs-137 activities along slopes Along three slope sections where the experiments reported in Kuhn and Yair (2003) and Kuhn et al. (2004) were conducted, the regolith profile was sampled for an analysis of Cs-137 activities. The activity of the Cs-137 was considered to be a tracer for infiltration and colluvial deposition or, in case of relatively low activities, an indication for prevailing runoff and erosion. Sampling and sample preparation followed the protocol outlined in Zapata (2002). Samples were taken in depth intervals of 10 cm down to the bedrock. They were then oven dried at 105 °C to evaporate any residual moisture, gently disaggregated using a rubber-headed pestle and mortar, passed through a 2 mm sieve and then weighed out each to ca. 100 g (+− 0.1 g). Each sample was then be placed in a small cylindrical plastic pot ready for counting lasting 24 h ca. 86,000 counts) which was sufficient to obtain results from Cs-137 to a precision of between +− 5 and 10%. Based on the peak area of the counts, the Cs-137 activity was calculated (Quine and Walling, 1993). The Cs-137 activities provide evidence that this pattern of infiltration, erosion and deposition occurs regularly (Fig. 5). On the ridges, the activities are lowest and Cs-137 cannot be found below the immediate crusted layer at the surface. Further down the slope where the regolith layer becomes thicker, the Cs-137 activities increase and reach further into the soil profile. The deepest activities were measured at approximately 50 cm. It is unlikely that this depth represents the thickness of a layer deposited since the first Cs-137 was released into the atmosphere during the 1950s and 1960s. More likely, the Cs-137 has moved through pores and cracks to this depth (Helalia et al., 1988; Parsons and Foster, 2011). The Cs-137 activity pattern along the slopes indicates that the ridges are lowered by rainflow erosion, but that the slopes do not retreat. Export of sediment from the straight sideslopes occurs as a consequence of undercutting and subsequent mass wasting of the lower sideslopes during runoff in the main channel (Fig. 6). 4.2. Erosion, deposition and badland topography The runoff and erosion data, as well as the properties of the surface material presented above are limited in two ways: first, they cover only Table 1 Rill system properties observed by Kuhn et al. (2004). The differences in regolith depth and averagerill width are significant (t-test of means, POE = 0.05). Regolith Difference in regolith Average rill width depth between top depth and bottom of slope (cm) (cm) (cm) Single rills 22 Integrated rill system 15 on concave slope
12 1
3.1 4.6
Difference in rill width between bottom and top of slope (cm) 1.4 3.5
7
a small proportion of the badlands, in most cases they were derived from just a slope; and second, while suggesting that runoff discontinuity and deposition explain the discrepancy between short-term erosion rates and long-term denudation, they do not illustrate the development of the current badland topography over the past 45,000 years. However, the topography of the Zin Valley Badlands provides further evidence for the significance of the rainfall-induced spatial separation between actively eroding and dormant slopes on landscape development. Most valleys are narrow, but long, suggesting that headward erosion dominates, controlled by concentrated flow in integrated rill networks and the valley channel (Fig. 2). The greater frequency of runoff in integrated rill networks at valley heads, the deposition of sediment eroded on the ridges on straight sideslopes, and small slides in the lower part of the valley sideslopes link the morphology to current day erosion processes. Using DEMs for both the entire Zin Valley badlands as well as a single slope section, this qualitative interpretation was tested. The analysis of a high resolution DEM for a slope section aimed at providing spatial information on the applicability of the hydrological and geomorphologic processes identified during the erosion tests, rill and regolith mapping. Micro-scale topography and rill heads of a hillslope section were surveyed using a Topcon IS03 imaging station. The automated total station acquires point clouds with very high accuracy (±5 mm). The data were linearly interpolated to a gridded DEM with 0.02 m resolution. DEM analysis was conducted using TopoToolbox (Schwanghart and Kuhn, 2010). Fig. 12 shows a high-resolution DEM of the scanned hillslope section. Deeply incised, continuous rills are found in planform concave sections while rills on convex parts of the hillslope are shallow and discontinuous. These patterns are reflected by the rill and slope profiles (inset of Fig. 12). On the concave upper slopes (profile 1), rills initiate close to the divide and follow a longitudinal profile that is rather straight along the entire slope. On the planform convex upper slopes (profile 2), rills form further away from the divide and often terminate in a sharp concave section in the middle of the slope. The lower convex part is partly incised by short rills. As a consequence of the topography, the area contributing runoff to the rills is very limited (often less than 1 m2) and located at the headwater area of the rills, with no runoff contribution from the interfluves separating two adjacent rills. As a consequence, water transmission losses are more pronounced in rills on planform convex slopes due to the lack of rill network integration. Rills in concave hillslope sections, on the other end, tend to collect more water through the capture of neighboring rills and their contributing areas, leading to downslope rill growth and continuity and further incision of the microcatchment. The lower regolith development in integrated rill networks (Table 1) reflects such greater erosion than on straight slopes with single rills. Overall, the high resolution DEM indicates that the spatial patterns of runoff generation, integration and continuity, associated with similar patterns of erosion and deposition, have an effect on slope profiles in the Zin Valley badlands. A regional scale DEM for the eastern part of the badlands was used to assess the effect of pronounced pattern of erosion on concave slope sections and deposition on the middle and lower sections of straight slopes. A DEM generated from contour lines with 5 m equidistance using the TopoToRaster interpolation tool in ArcGIS 10 (Hutchinson, 1989). Sub-basin outlets were mapped along the Havarim channel (Fig. 13). Channel heads were mapped by visual image interpretation in Google Earth. The elongation ratio (Re) of the basins can be used as an index for the relevance of headward erosion. Re describes the ratio between the diameter of a circle with the same area as the basin and the maximum length of the basin parallel to the principal drainage line (Schumm, 1956). The majority of the basins have elongation ratios (Fig. 14) of less than 0.55, which indicates that the basin maximum length is about twice the diameter of a circle of the same area. Low values of Re have been interpreted to reflect a predominant headward growth of the drainage catchments (Schumm, 1956).
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
8
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 12. Slope-scale DEM of slope with integrated rill (1) network and single rill (2). The effect of runoff continuity and erosion in the integrated rill network and the runoff discontinuity and deposition have led to the development of different slope profiles (inset top left corner). The scanned slope is located at 30.84° N and 34.75°.
5. Resilience of Zin Valley badlands to climate change The results of erosion tests, regolith mapping and sampling and the DEM analysis reviewed above show that a link between current surface processes, form and landscape development in the Zin Valley badlands is plausible. The data also reconcile the high erosion rates observed on (some) slope sections with the age of the badlands. The key to the improved understanding of the geomorphology of the badlands is a holistic perspective that more than 40 years of research offer to generate an understanding of the spatial patterns of surface characteristic, geomorphic processes, their current rates and
the role of rainfall for runoff and erosion. The spatial differentiation of erosion and deposition processes within a watershed of uniform lithology also questions the close relationship between lithology and surface processes assumed for badlands. Surface erosion by water and mass wasting of slopes are the two basic modes for formation of badlands (e.g. Campbell, 1997; Howard, 1994; Wainwright and Brazier, 2011). The prevalence of either process domain is generally attributed to differences in lithology, most notably water erosion for sandstones and mass wasting for shales. The occurrence of both forms of erosion in the Zin Valley badlands questions this simple conceptual model. The prevalence of runoff discontinuity on large slope sections for
Fig. 13. Regional scale DEM of the western portion of the Zin Valley Badlands drained by the Navarim contributory. Black lines mark the first order catchments and drainage lines.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 14. Elongation ratio analysis (Re) extracted from the 5 m regional DEM of the western section of the Zin Valley Badlands.
current rainfall conditions, and the corresponding differentiation of surface characteristics, highlights the role of rainfall as a controlling factor for badland form and process. Consequently, rainfall characteristics appear to control current landscape development in the Zin Valley Badlands. However, this control is not linear, determining only the rate of erosion, as suggested by the model landscape approach, but strongly influenced by the geomorphic history of the badlands. In particular, the formation of colluvial deposits acting as a runoff sink fosters an “autostabilisation” of the slopes, increasing their resilience to climate change. The postulated positive feedback between discontinuous runoff, formation of a colluvium and slope infiltration capacity sheds a critical light on the assumption that up to 300 mm of annual rainfall were required to incise the badlands, limiting their formation to more humid periods (Yair et al., 1980a, 1980b), common in the Levante during the Pleistocene (Issar, 2001). On shorter, juvenile slopes without a colluvium, runoff continuity would be improved, even under current rainfall characteristics (Kuhn, 2011), providing a mechanism for the evacuation of the eroded sediment. In turn, the presence of the colluvium today requires a much greater change in rainfall magnitude to overcome the colluvial runoff sink and increase slope erosion than during the late Pleistocene when incision started on no colluvium had formed. In addition, the frequency of individual rainfall events with a magnitude sufficient for continuous runoff, erosion and sediment export out of first order basins is more important than the annual rainfall amount.
9
The results presented above reconcile the contradiction between high local erosion rates and the age of the badlands by showing that large parts experience little change and are also resilient to change. Thereby, they confirm that the Zin Valley Badlands are indeed not simple landscapes with immediate reactions to climate change. The patterns of the surface characteristics determining infiltration, runoff, erosion and deposition also offer an explanation for the variability of erosional responses observed in other badlands. Basically, one can speculate that without an understanding of the identified scalerelated issues most observations do not deliver data that can be compared between badlands. To do so, slope sections and catchments with similar relevant process domains would have to be observed, rather than plots of a particular size on slope sections with many erosion features. The results discussed above clearly indicate that badlands are more complex than the classic geomorphic literature suggests. However, we would argue that badlands can still be considered as model landscapes. Our main arguments are that despite the spatial heterogeneity of surface properties and processes, badlands are still “simpler” than other landscapes where vegetation and land use modify surface processes. In particular, slopes and fluvial valleys altered by humans represent geomorphic systems in disequilibrium with the driving erosive forces and may thus be rendered particularly sensitive to climate change and high magnitude events. The patchiness of erosion and deposition in badlands also offers conceptual guidance for further questions to be addressed by geomorphologists. For example, the development of a colluvium would lead to a runoff sink due to the increased infiltration capacity of the deposited material compared to less porous and permeable bedrock in any climate. As a consequence, the response to climate change of such a system is muted unless the colluvial buffer is saturated with water or eroded away. So while badlands are more complex than perceived, most landscapes are probably also more complex than we think. Acknowledgments Our thanks go to three generations of researchers working in the Zin Valley badlands. For the most recent work, we would like to acknowledge the support of James Grapes of the Isotope Laboratory at the University of Exeter and the Freie Akademische Gesellschaft for sponsoring the field visit of Nikolaus J. Kuhn and Wolfgang Schwanghart in 2008. References
6. Conclusions The review and results presented above show that the resilience of the Zin Valley Badlands to climate change is relative, depending on the spatial and temporal scale applied. Runoff generation on the ridges is very sensitive to changes in rainfall magnitude, more or less producing a linear positive response to an increase in effective rainfall above the low threshold of runoff generation. The colluvial runoff sink that has formed as a consequence of runoff discontinuity, on the other hand, buffers runoff contribution from slopes to first order channels. As a consequence, a non-linear reaction of slope runoff is likely in response to increasing rainfall. Initially, the frequency of continuous runoff events would increase with more rainfall events above the critical threshold for runoff generation. One can speculate that this would also lead to (i) less deposition and over time even an erosion of the colluvium and (ii) an increase of the extent of rapidly sealing thin crust currently present only at the ridges. These changes in surface characteristics associated with the removal of the colluvial runoff sink would lower the runoff effective rainfall required for continuous runoff from slopes, increasing main channel runoff, sediment export and eventually incision and thus, at least down to the local base level of erosion, rejuvenate the badland formation.
Arkin, Y., Nathan, Y., Starinsky, A., 1972. Paleocene–Early Eocene environments of deposition in the northern Negev. Bulletin — Geological Survey of Israel 56, 1–17. Bentor, Y.K., 1966. The Clays of Israel. Guidebook to the Excursions. The International Clay Conference, Jerusalem, Israel, p. 123. Branson, F.A., Gifford, G.F., Renard, K.G., Hadley, R.F., 1981. Rangeland Hydrology, Range Science Series 1. Society for Range Management, Denver, CO. Bryan, R.B., Yair, A., 1982. Perspectives of studies of badland geomorphology. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich,CT, pp. 1–12. Campbell, I.A., 1997. Badlands and badland gullies. In: Thomas, D.S.G. (Ed.), Arid Zone Geomorphology: Process, Form and Change in Drylands. Belhaven Press, London, Halsted Press, New York, pp. 261–291. Della, Seta M., Del Monte, M., Fredi, P., Palmieri, E.L., 2007. Direct and indirect evaluation of denudation rates in Central Italy. Catena 71, 21–30. Fox, N.I., 2004. Technical Note: the representation of rainfall drop-size distribution and kinetic energy. Hydrology and Earth System Sciences 8 (5), 1001–1007. Gallart, F., Llorens, P., Latron, J., Regüés, D., 2002. Hydrological processes and their seasonal controls in a small Mediterranean mountain catchment in the Pyrenees. Hydrology and Earth Systems Sciences 6 (3), 527–537. Helalia, A.M., Letey, J., Graham, R.C., 1988. Crust formation and clay migration effects in infiltration rate. Soil Science Society of America Journal 52, 251–255. Hodges, W.K., Bryan, R.B., 1982. The influence of material behaviour on runoff initiation in the Dinosaur Badlands, Canada. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich, CT, pp. 13–47. Howard, A.D., 1994. Badlands. In: Abrahams, A.D., Parsons, J.P. (Eds.), Geomorphology of Desert Environments. Chapman & Hall, London, pp. 213–242. Hutchinson, M.F., 1989. A new procedure for gridding elevation and stream line data with automatic removal of spurious pits. Journal of Hydrology 1989 (106), 211–232.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
10
A. Yair et al. / Catena xxx (2012) xxx–xxx
Issar, Arie S., 2001. Paleo-environments of the uppermost Quaternary in Israel. Current Research, GSI 12, 235–238. Jenny, H., 1941. The Factors of Soil Formation. McGraw-Hill, New York. Kinnell, P.I.A., 2009. The influence of raindrop induced saltation on particle size distributions in sediment discharged by rain-impacted flow on planar surfaces. Catena 78, 2–11. Kuhn, N.J., 2011. Rainfall, runoff and environmental change in the critical zone of two dryland environments. GEOÖKO XXXII, 21–40. Kuhn, N.J., Yair, A., 2003. Spatial distribution of surface conditions and runoff generation in small arid watersheds, Zin Valley Badlands, Israel. Geomorphology 57, 183–200. Kuhn, N.J., Yair, A., Kasanin-Grubin, M., 2004. Spatial distribution of surface properties, runoff generation and landscape development in the Zin Valley Badlands, northern Negev, Israel. Earth Surface Processes and Landforms 29. Kutiel, H., 1978. The distribution of rain intensities in Israel. MSc Thesis, The Hebrew University, Jerusalem. Marks, A.E., 1977. A preliminary overview of central Negev Prehistory, Israel, vol. II. SMU Press, Dallas, pp. 3–31. Morin, J., Cluff, B.C., Powers, W.R., 1970. Realistic rainfall simulation for field investigations. Paper H78. 51st Annual Meeting, American Geophysical Union, Washington, DC. Nadal-Romero, E., Martínez-Murillo, J.F., Vanmaercke, M., Poesen, J., 2011. Scale dependency of sediment yield from badland areas in Mediterranean environments. Progress in Physical Geography June 2011, vol. 35 no. 3, pp. 297–332. Nathan, Y., 1978. Studies on Palygorskite. PH.D.thesis, The Hebrew University of Jerusalem pp 156 (in Hebrew) . Parsons, A.J., Foster, I.D.L., 2011. What can we learn about soil erosion from the use of 137Cs? Earth-Science Reviews 108, 101–113.
Quine, T.A., Walling, D.E., 1993. Use of Caesium-137 measurements to investigate relationships between erosion rates and topography. In: Thomas, D.S.G., Allison, R.J. (Eds.), Landscape Sensitivity. : BGRG Symposia Series. John Wiley and Sons, Chichester, pp. 31–48. Schumm, S.A., 1956. The role of creep and rainwash on the retreat of badland slopes. American Journal of Soil Science 254, 693–706. Schwanghart, W., Kuhn, N.J., 2010. TopoToolbox: a set of Matlab functions for topographic analysis. Environmental Modelling & Software 25, 770–781. Torri, D., Colica, A., Rockwell, D., 1994. Preliminary study of the erosion mechanisms in a biancana badland (Tuscany, Italy). Catena 23, 281–294. Torri, D., Calzolari, C., Rodolfi, G., 2000. Badlands in changing environments: an introduction. Catena 40, 119–125. Wainwright, J., Brazier, R.E., 2011. Slope systems. In: Thomas, D.S.G. (Ed.), Arid Zone Geomorphology: Process, Form and Change in Drylands. John Wiley and Sons, Ltd. Yair, A., Kossovsky, A., 2002. Climate and surface properties: hydrological response of small arid and semi-arid watersheds. Geomorphology 42, 43–57. Yair, A., Lavee, H., Bryan, R.B., Adar, E., 1980a. Runoff and erosion process and rates in the Zin Valley badlands, Northern Negev, Israel. Earth Surface Processes 5, 205–225. Yair, A., Lavee, H., Goldberg, P., Bryan, R.B., 1980b. Present and past geomorphic evidences in the development of a badlands landscapes, Zin valley Badlands, northern. Yair, A., Goldberg, P., Brimer, B., 1982. Long term denudation rates in the Zin–Havarim badlands, northern Negev, Israel. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich, CT, pp. 279–292. Zapata, F. (Ed.), 2002. Handbook for the assessment of soil erosion and sedimentation using environmental radionuclides. : The chapter in question is Chapter 4. Kluwer Acdemic Publishers, London, pp. 59–65.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
Contents lists available at SciVerse ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
The resilience of a badland area to climate change in an arid environment Aaron Yair a, Rorke B. Bryan b, Hanoch Lavee c, Wolfgang Schwanghart d, Nikolaus J. Kuhn e,⁎ a
Hebrew University Jerusalem, Israel University of Toronto, Canada c Bar Ilan University, Tel Aviv, Israel d University of Potsdam, Germany e University of Basel, Switzerland b
a r t i c l e
i n f o
Article history: Received 17 February 2012 Received in revised form 5 April 2012 Accepted 15 April 2012 Available online xxxx Keywords: Badlands Model landscape Climate change Resilience
a b s t r a c t Badlands have long been considered as model landscapes due to their perceived close relationship between form and process. The often intense features of erosion have also attracted many geomorphologists because the associated high rates of erosion appeared to offer the opportunity for studying surface processes and the resulting forms. Recently, the perceived simplicity of badlands has been questioned because the expected relationships between driving forces for erosion and the resulting sediment yield could not be observed. Further, a high variability in erosion and sediment yield has been observed across scales. Finally, denudation based on currently observed erosion rates would have lead to the destruction of most badlands a long time ago. While the perceived simplicity of badlands has sparked a disproportional (compared to the land surface they cover) amount of research, our increasing amount of information has not necessarily increased our understanding of badlands in equal terms. Overall, badlands appear to be more complex than initially assumed. In this paper, we review 40 years of research in the Zin Valley Badlands in Israel to reconcile some of the conflicting results observed there and develop a perspective on the function of badlands as model landscapes. While the data collected in the Zin Valley clearly confirm that spatial and temporal patterns of geomorphic processes and their interaction with topography and surface properties have to be understood, we still conclude that the process of realizing complexity in the “simple” badlands has a model function both for our understanding as well as perspective on all landscape systems. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Badlands are often considered as an intensively dissected barren landscape, devoid of soil cover, with a sparse or absent vegetation (e.g. Howard, 1994). The very high drainage densities, V-shaped valleys and steep slopes, have often led to believe that badlands represent a landscape where the frequency and magnitude of overland flow and erosion processes are high to very high, resulting in rapid landscape evolution (see review by Wainwright and Brazier, 2011). This is why badlands have been sometimes regarded as “ideal field laboratories” for testing landscape evolution hypothesis (Bryan and Yair, 1982; Campbell, 1997; Schumm, 1956). Recent research has shed a critical light on the “simple landscape” perspective. Two key issues support this change in perspective. First, the erosion rates and sediment yields observed in badland areas vary widely. A recent comprehensive review, conducted by Nadal-Romero et al. (2011) of badland areas in Mediterranean environments, exhibits a huge variability in erosion rates. No relationship in erosion
⁎ Corresponding author. E-mail address: [email protected] (N.J. Kuhn).
rates was found for areas in the range of 0.0001–100 ha. What is more, erosion rates for areas as small as 0.1 m2 vary in the range of 1.1– 500 kg m− 2y− 1. A more pronounced variability (0.12–1000 t ha− 1y− 1) was obtained for larger areas covering 10–100 ha. For areas in the range of 100 ha to 10 km2 erosion rates are smaller, but still a ten-fold variability exists. The observed differences from hillslope to catchment scale are at least on the same order of magnitude than those typically observed in a wide range of drylands (reviewed in Branson et al., 1981). They are, apart from differences in lithology and topography, attributed to different patterns of sediment transport and deposition in the studied badlands and the scale effects of erosion and deposition (Nadal-Romero et al., 2011). No relationship was found between average annual rainfall and average erosion rates for annual rainfalls in the wide range of 80–1250 mm (Nadal-Romero et al., 2011). The limited effect of rainfall magnitude was also identified by Della et al. (2007), who stated that the threshold for runoff generation is 60 mm for 5 consecutive days. Below this threshold, surface runoff does not seem to reach the power required for significant modifications of the topographic surface. The observations in the studies cited above confirm data presented in comprehensive and detailed earlier papers (Della et al., 2007 and Gallart et al., 2002). Overall, the variability in sediment yields and limited relationship with annual rainfall indicate that the badlands are by no means simpler than more complex landscapes.
0341-8162/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.catena.2012.04.006
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
2
A. Yair et al. / Catena xxx (2012) xxx–xxx
A key issue limiting the use of badlands as simple, model landscapes appears to be the very high spatial variability in erosion rates observed under present day rainfall conditions. The spatial variability presents a serious barrier for the extrapolation of erosion rates from the hill slope to catchment scales and certainly to landscape evolution over longer geological time scales under changing climatic conditions. A second reason contradicting the simple geomorphic landscape is the mismatch observed between current erosion rates and the longterm denudation rates that have been reconstructed for several badland areas. For example, denudation rates observed by Yair et al. (1982) in the Zin Valley Badlands were on the order of 2–5 cm y − 1 and would be expected to completely wipe out an area with a relative elevation of 10 m within the very short time of 200 to 500 years. Similar mismatches have been reported for the Dinosaur badlands in Alberta (Hodges and Bryan, 1982) and to a lesser extent badlands in Tuscany (Torri et al., 1994). This is quite unreasonable unless we assume those denudation rates equal extremely high rates of tectonic uplift, or alternatively that high erosion rates result from extreme rain events with an extremely high frequency. A final reason for the presumed geomorphic simplicity of badlands may be their perception as a model landscape, which led to ignoring the complexity and spatial variability of surface processes in badland studies. Without singling out an individual study (for a selection of the methods used in badland research see Bryan and Yair (1982) and Torri et al. (2000)), we would just like to draw the attention to the way erosion was measured and extrapolated. Often runoff and sediment yield were monitored on plots and small catchments of varying size, often equating sediment yield at the lower end of the plot with erosion and thus ignoring deposition on the plot. This must invariably lead to high net erosion rates when plots are small compared to the average transport distance of the detached and entrained material (Kinnell, 2009). This problem is exacerbated because the selection of monitoring sites is often guided by the presence or absence of erosion features, without assessing their spatial representation for the entire badlands. In the light of this critical assessment of the idea of badlands as a “natural laboratory” for landscape development, 40 years of research activities in the Zin Valley badlands in the Negev desert are reviewed in this paper. The wide range of studies on different spatial and temporal scales illustrates the gradual development of our understanding of badlands and suggests some answers to some conflicting results mentioned above. The paper is split into two sections: first, the published work is reviewed, followed by some previously unpublished data acquired in the past five years focusing on linking processes to landscape development. 2. Study area The Zin Valley Badlands are located in the northern Negev desert (Fig. 1). They are incised into the clay rich shales of the Taqyia formation of Paleocene age. The field studies discussed in this review were conducted at the lower part of this formation where the clay fraction is 65–75% of the bedrock (Bentor, 1966). The clay mineralogy is dominated by 50% of Montmorillonite, 40% Kaolinite and 10% Illite (Arkin et al., 1972; Nathan, 1978). The present topography consists of generally V-shaped longitudinal valleys, with rounded ridges along the divides, straight sideslopes of approximately 30° angle, and rillto gully sized fluvial channels along the valley floor (Fig. 2). Average annual rainfall in the region, observed over 50 years, is 90 mm per year with extreme values of 34 and 170 mm, potential annual evaporation is 2500 mm (Yair and Kossovsky, 2002). Only few rainstorms yield more than 20 mm per day. Two well-defined terraces (Fig. 3) with an extensive scatter of prehistoric sites pointing at an intensive activity at the late Quaternary have been identified in the area (Yair et al., 1982). This is attributed to the extensive outcrop of the Mishash formation, which is rich in flint used for the preparation of working tools, and
Fig. 1. Location and average annual precipitation at the Zin Valley Badlands in the Negev desert, Israel.
the occurrence of nearby perennial springs along the Zin river. The lower terrace is situated below 400 m. Prehistoric sites here are attributed to the Upper Paleolithic period which span an interval of 45,000 to 25,000 years BP. The upper terrace (above 400 m) is composed of colluvial silts containing in situ Epi-Paleolithic sites dated by C 14 to 13,530 BP (Marks, 1977). The incision between the terraces ranges between 5 and 7 m, equalling average annual channel incision rates of 0.25–0.35 mm. These incision rates are an order of magnitude below those calculated based on erosion tests conducted in the past decades (see review of studies below) and illustrate that the conflicting results between short and long-term studies identified in the Introduction also apply to the Zin Valley Badlands. The prehistoric findings also clearly indicate that the main phase of erosion took place during the Upper Pleistocene when the level of the Dead Sea lake was approximately 200 m above its present level, representing a wet climatic phase, leading to the assumption that about 300 mm of annual rainfall were required to incise the badlands (Yair et al., 1980b). This adds a further complication to the interpretation of the data collected
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
3
N
Fig. 2. Aerial photograph of the western section of the Zin Valley Badlands. Note the longitudinal form of the badlands, indicating a relatively high rate of headward erosion. The size of the image is approximately 0.25 km2.
on current erosion rates in the Zin Valley Badlands: if erosion was higher during a phase of wet climate in the past, why haven't the badlands been eroded away completely a long time age if current erosion rates are still so high? 3. Surface characteristics, infiltration and erosion
Fig. 4. Crusts on north-(left) and south-facing slopes. Note the rough popcorn structure on the left and the smooth microtopography on the right.
3.1. Surface characteristics The slopes in the Zin Valley Badlands show roughly three distinct sets of surface characteristics (Figs. 4 and 5). On the ridges, a poorly developed regolith dominates. Above a layer of shards formed of physically weathered bedrock, a thin, 2 to 5 mm thick crust, broken by desiccation cracks, has formed. On north-facing sideslopes, a 20 to 30 cm deep, crusted soil with deep desiccation cracks and a “popcorn” surface microtopography with depressions of 4 to 8 cm depth
has developed (Yair et al., 1980a). On these slopes, seals of lichens are often found. Regolith on south-facing slopes is less developed than on the north-facing slopes, reaching a thickness of only 10 cm. Three distinctly different rill systems have developed on straight valley sideslopes and along concave valley heads (Figs. 4 and 6). On straight and plan north-facing slopes the surface is generally unrilled with the exception of desiccation cracks where small pipes can develop (Yair et al., 1982). On south-facing slopes, single, discontinuous rills dominate. Well-integrated rill systems are common along valley heads or concave slope segments where the topography of the slopes enhances the convergence of rills and the formation of rill networks. It is noteworthy that most rills begin near the ridges and do show only limited change in width and depth along the slope. 3.2. Small-scale erosion tests
Fig. 3. Topography and sites with Paleolithic artefacts on the upper and lower terrace of the Navarim contributory (dark line in upper left part of figure) to the Zin River (crossing S–N on right of figure).
Runoff and erosion on the different interrill and rill features have been studied using a range of infiltration and erosion tests. Two sets of sprinkling experiments were carried out. The first was conducted in 1979 (Yair et al., 1980a, 1982) and the second in 2001 (Kuhn and Yair, 2003; Kuhn et al., 2004). The first set was carried out on two different scales: small plots covering 1.5 m 2 and larger plots covering up to 45 m 2. The use of different plot sizes aimed at identifying potential scale effects on runoff and erosion. The sprinkling experiments on the small plots were conducted on opposite north- and south-facing slopes. Sprinkling on the small plots was conducted with the Morin rainfall simulation unit (Morin et al., 1970). This unit (Fig. 7) can reproduce accurately the intensity and kinetic energy characteristics of natural rainfall. Rainfall was applied at 35.3 mm h− 1 for 43 min, totalling 25.3 mm (Yair et al., 1980b). Runoff on the north-facing slope started after 5 min and increased gradually (Fig. 8). Total runoff yield was
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
4
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 5. Crust properties, regolith depth and Cs-137 activities along a Northwestern-facing slope section studied by Kuhn and Yair (2003).
11.98 mm representing a runoff coefficient of 47.4%. However, infiltration rate after 10 min was still quite high (27 mm h− 1), and 18 mm h− 1 after 20 min. Despite the high rain intensity no surface disturbance by splash was observed and no change in surface microtopography took place. Sediment yield was 667 g m − 2, representing a denudation rate of 0.48 mm, which must be considered as quite low for the high rain intensities and long durations applied. Rainfall at the south-facing plot was applied at 36.7 mm h− 1 and lasted 48 min, totalling 29.4 mm. Runoff started after 11 min (Fig. 8), increased gradually, and approached equilibrium conditions at minute 37. Runoff yield was 8.64 mm representing a runoff coefficient of
29.4%, much lower than that of the opposite slope. Infiltration rate, after 10 and 20 min, was still very high (37 mm h− 1 and 20 mm h− 1, respectively). Again, the change in surface characteristics during the 48-minute duration of the test was negligible. Sediment yield was low compared to the north-facing slope (38 g m − 2), representing a denudation rate of 0.027 mm. It is interesting to note that runoff and erosion on the north-facing slopes were greater than on those facing south, despite the much greater surface roughness and depression storage on the north-facing slopes. The greater infiltration on the south-facing slopes was attributed by Yair et al. (1980b) to the thinner, more porous crusts and shards underlying the thin top layer (Fig. 5). 3.3. Infiltration on crusts along a slope
Fig. 6. Rill systems and headward erosion on N-facing slopes in the Zin Valley Badlands. Straight and radial slopes show little or no rill development while on concave slopes integrated rill networks dominate. Undercutting by the main channel (lower right of image center) indicates that colluvial sediment is transferred into the main channel by mass wasting.
A second set of small-scale measurements were conducted on plots across slopes by Kuhn and Yair (2003) and Kuhn et al. (2004) using a needle sprinkler covering an area of 1 m 2 (Fig. 7). The tests aimed at bridging a gap in the previous observations by focusing on the effect of different crust characteristics observed on one slope on infiltration during quasi-natural rainfall event. The simulator produced drops of 3.5 mm diameter and a kinetic energy of 3.1 J m− 2 mm− 1 (Kutiel, 1978). These values are lower than for natural rainfall (Fox, 2004), but the existence and stability of a crust when rained on with the Morin unit described above indicates that this does not affect infiltration. Low intensity rainfall was applied for 15 min, followed by a high intensity rain burst (35 mm h− 1) for 10–12 min. This pattern simulates a natural rainstorm, when a high intensity rain shower occurs during a rainfall event with already wetted soil. The return interval of such an event is approximately 5 years in the Sede Boqer region (Yair and Kossovsky, 2002). Infiltration differed significantly for the different crust types (Fig. 9). Highest infiltration rates were obtained for the valley side slopes and the rills formed there (0.58 mm min− 1, not shown in Fig. 9 because only a final rate was observed). The ridges exhibited much lower infiltration and more
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
5
Fig. 7. The Morin (left) and Kuhn and Yair (right) rainfall simulator in the Zin Valley Badlands.
rapid runoff development than the side slopes (Fig. 9), indicating that partial runoff is generated during rainfall events of high frequency and low magnitude.
3.4. Slope-scale infiltration and erosion tests Slope‐scale infiltration tests were conducted using a sprinkler, which provided uniform coverage of 90 m 2 (Yair et al., 1982), but which does not accurately reproduce natural velocities of rainfall. The drop‐size spectrum is smaller than for natural rainfall and the fall height of the drops generates only 50% of the natural kinetic energy. However, like with the sprinkling tests conducted by Kuhn and Yair (2003), the existence of a stable crust indicated that the limited kinetic energy did not affect infiltration. The north-facing plot covered an area of 45 m 2, but runoff was collected only from the central 30 m 2. Average rain intensity applied over the sprinkled area was 48.5 mm h− 1. The test lasted 34.5 min, representing a rain amount of 28 mm. Runoff developed only in rills and displayed a pulsating character and varied in space and time (Fig. 10). Pipe flow was observed after 15 min and after 30 min rill bank and pipe collapse was widespread. The intermittency of the runoff is attributed to clogging of rills, either by the collapse of banks or deposition of sediment once rill infiltration losses reduce transport capacity of flow below the actual load. Similarly, collapsing pipes diverted runoff onto interrill areas where infiltration losses were high (Fig. 11). Runoff yield was 3.94 mm, representing a runoff coefficient of 14.2%, much lower than that obtained in the small plot. Due to the pulsating character of runoff, sediment concentration varied in the range of 32–172 g l− 1. Sediment yield was 235 g m− 2, representing an average denudation rate of 0.17 mm. All of the sediment originated from the rill banks and pipes along the rills. The south-facing plot covered also an area of 45 m 2, but runoff was collected from the central 30 m 2. Average rain intensity was only 18 mm h− 1. The test lasted 51.5 min (Fig. 10). Runoff started, on the ridge, after 5.5 min, and was quickly absorbed in desiccation cracks that did not seal. Continuous flow was limited to the rills, but did not occur simultaneously in all rills. The intermittent runoff generated only 9 l of discharge into the collecting tank, which corresponds to a runoff coefficient of only 1.7%. Sediment discharge was also very low (85 g m − 2), representing a denudation rate of 0.06 mm. 3.5. Surface characteristics and spatial patterns of runoff generation and continuity
Fig. 8. Runoff observed by Yair et al. (1980a) on small1.5 m2 plots with different aspect in the Zin Valley Badlands.
The infiltration tests show that runoff generation in the Zin Valley Badlands is highly variable in space and time. Runoff is generated more frequently on the ridges (30% of annual rainfall, than on the valley sideslopes (10% of annual rainfall) (Kuhn and Yair, 2003). This pattern leads to runoff discontinuity. A key question for understanding the complexity of badlands as a hydrologic and geomorphic system is the long-term interaction between runoff and infiltration patterns, the corresponding feedback on surface properties, in particular erosion
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
6
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 9. Infiltration rates observed by Kuhn and Yair (2003) during simulated rainfall on ridge and slope interrill areas in the Zin Valley Badlands. The error bars mark the maximum deviation from the mean observed during three replicate tests.
and deposition, and the consequences for landscape development. In the second part of this paper, an analysis of surface properties and topography aims at addressing the interaction between form and process and the long-term landscape development. The low infiltration on the ridges leads to rainflow erosion and limits soil development, maintaining soil susceptible to rapid sealing and thus creating a positive feedback on low infiltration (Kuhn and Yair, 2003). On the sideslopes, infiltration of rainfall and water flowing downslope from the ridges, cause flow discontinuity and deposition, which enhances soil development, leading to high infiltration and discontinuous runoff. The long-term effects of runoff generation and frequent discontinuous flow are reflected by soil development along the badland slopes. Soil depth is indicative of the balance between erosion and soil formation, being deeper in areas of low erosion, limited runoff and high infiltration, and shallow in areas with much runoff and erosion (Jenny, 1941). In the Zin Valley Badlands, weathering depth is
lowest at or close to the divide (Fig. 5). Downslope, weathering depth increases initially for 4 to 5 m, but declines again in the lower part of the slope. There is also a clear difference in weathering depth between non-rilled slope sections and those with an integrated rill system (Table 1). In the integrated rill system, weathering depth is generally shallower than in the single rill and almost constant along the entire
Fig. 10. Runoff observed by Yair et al. (1982) during slope-scale erosion tests in the Zin Valley Badlands.
Fig. 11. Discontinuous runoff, marked by disappearing dark dye tracer pointed out by white arrows, on north-facing slope during erosion test conducted by Yair et al. (1982).
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
slope, showing that, due to more frequent runoff, erosion rates are higher as well. Kuhn and Yair (2003) concluded that the frequent discontinuity of surface runoff is responsible for this pattern. Using the data from their infiltration tests, they estimated that rainfall intensities of as low as 10 mm per hour are sufficient to overcome infiltration losses in an integrated rill network while at least 25 mm are required for interrill areas. Dye tracers placed on ridges, interrill areas and in rills to monitor runoff during natural rainstorms support the results of the sprinkling tests when rainfall of 26 mm on December 6th 2001 was sufficient to produce continuous flow from ridges to valleys. However, even during this storm no surface runoff from the interrill areas on the sideslopes reached the rills (Kuhn and Yair, 2003). 4. Effect of partial area runoff generation and discontinuity on landscape development 4.1. Cs-137 activities along slopes Along three slope sections where the experiments reported in Kuhn and Yair (2003) and Kuhn et al. (2004) were conducted, the regolith profile was sampled for an analysis of Cs-137 activities. The activity of the Cs-137 was considered to be a tracer for infiltration and colluvial deposition or, in case of relatively low activities, an indication for prevailing runoff and erosion. Sampling and sample preparation followed the protocol outlined in Zapata (2002). Samples were taken in depth intervals of 10 cm down to the bedrock. They were then oven dried at 105 °C to evaporate any residual moisture, gently disaggregated using a rubber-headed pestle and mortar, passed through a 2 mm sieve and then weighed out each to ca. 100 g (+− 0.1 g). Each sample was then be placed in a small cylindrical plastic pot ready for counting lasting 24 h ca. 86,000 counts) which was sufficient to obtain results from Cs-137 to a precision of between +− 5 and 10%. Based on the peak area of the counts, the Cs-137 activity was calculated (Quine and Walling, 1993). The Cs-137 activities provide evidence that this pattern of infiltration, erosion and deposition occurs regularly (Fig. 5). On the ridges, the activities are lowest and Cs-137 cannot be found below the immediate crusted layer at the surface. Further down the slope where the regolith layer becomes thicker, the Cs-137 activities increase and reach further into the soil profile. The deepest activities were measured at approximately 50 cm. It is unlikely that this depth represents the thickness of a layer deposited since the first Cs-137 was released into the atmosphere during the 1950s and 1960s. More likely, the Cs-137 has moved through pores and cracks to this depth (Helalia et al., 1988; Parsons and Foster, 2011). The Cs-137 activity pattern along the slopes indicates that the ridges are lowered by rainflow erosion, but that the slopes do not retreat. Export of sediment from the straight sideslopes occurs as a consequence of undercutting and subsequent mass wasting of the lower sideslopes during runoff in the main channel (Fig. 6). 4.2. Erosion, deposition and badland topography The runoff and erosion data, as well as the properties of the surface material presented above are limited in two ways: first, they cover only Table 1 Rill system properties observed by Kuhn et al. (2004). The differences in regolith depth and averagerill width are significant (t-test of means, POE = 0.05). Regolith Difference in regolith Average rill width depth between top depth and bottom of slope (cm) (cm) (cm) Single rills 22 Integrated rill system 15 on concave slope
12 1
3.1 4.6
Difference in rill width between bottom and top of slope (cm) 1.4 3.5
7
a small proportion of the badlands, in most cases they were derived from just a slope; and second, while suggesting that runoff discontinuity and deposition explain the discrepancy between short-term erosion rates and long-term denudation, they do not illustrate the development of the current badland topography over the past 45,000 years. However, the topography of the Zin Valley Badlands provides further evidence for the significance of the rainfall-induced spatial separation between actively eroding and dormant slopes on landscape development. Most valleys are narrow, but long, suggesting that headward erosion dominates, controlled by concentrated flow in integrated rill networks and the valley channel (Fig. 2). The greater frequency of runoff in integrated rill networks at valley heads, the deposition of sediment eroded on the ridges on straight sideslopes, and small slides in the lower part of the valley sideslopes link the morphology to current day erosion processes. Using DEMs for both the entire Zin Valley badlands as well as a single slope section, this qualitative interpretation was tested. The analysis of a high resolution DEM for a slope section aimed at providing spatial information on the applicability of the hydrological and geomorphologic processes identified during the erosion tests, rill and regolith mapping. Micro-scale topography and rill heads of a hillslope section were surveyed using a Topcon IS03 imaging station. The automated total station acquires point clouds with very high accuracy (±5 mm). The data were linearly interpolated to a gridded DEM with 0.02 m resolution. DEM analysis was conducted using TopoToolbox (Schwanghart and Kuhn, 2010). Fig. 12 shows a high-resolution DEM of the scanned hillslope section. Deeply incised, continuous rills are found in planform concave sections while rills on convex parts of the hillslope are shallow and discontinuous. These patterns are reflected by the rill and slope profiles (inset of Fig. 12). On the concave upper slopes (profile 1), rills initiate close to the divide and follow a longitudinal profile that is rather straight along the entire slope. On the planform convex upper slopes (profile 2), rills form further away from the divide and often terminate in a sharp concave section in the middle of the slope. The lower convex part is partly incised by short rills. As a consequence of the topography, the area contributing runoff to the rills is very limited (often less than 1 m2) and located at the headwater area of the rills, with no runoff contribution from the interfluves separating two adjacent rills. As a consequence, water transmission losses are more pronounced in rills on planform convex slopes due to the lack of rill network integration. Rills in concave hillslope sections, on the other end, tend to collect more water through the capture of neighboring rills and their contributing areas, leading to downslope rill growth and continuity and further incision of the microcatchment. The lower regolith development in integrated rill networks (Table 1) reflects such greater erosion than on straight slopes with single rills. Overall, the high resolution DEM indicates that the spatial patterns of runoff generation, integration and continuity, associated with similar patterns of erosion and deposition, have an effect on slope profiles in the Zin Valley badlands. A regional scale DEM for the eastern part of the badlands was used to assess the effect of pronounced pattern of erosion on concave slope sections and deposition on the middle and lower sections of straight slopes. A DEM generated from contour lines with 5 m equidistance using the TopoToRaster interpolation tool in ArcGIS 10 (Hutchinson, 1989). Sub-basin outlets were mapped along the Havarim channel (Fig. 13). Channel heads were mapped by visual image interpretation in Google Earth. The elongation ratio (Re) of the basins can be used as an index for the relevance of headward erosion. Re describes the ratio between the diameter of a circle with the same area as the basin and the maximum length of the basin parallel to the principal drainage line (Schumm, 1956). The majority of the basins have elongation ratios (Fig. 14) of less than 0.55, which indicates that the basin maximum length is about twice the diameter of a circle of the same area. Low values of Re have been interpreted to reflect a predominant headward growth of the drainage catchments (Schumm, 1956).
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
8
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 12. Slope-scale DEM of slope with integrated rill (1) network and single rill (2). The effect of runoff continuity and erosion in the integrated rill network and the runoff discontinuity and deposition have led to the development of different slope profiles (inset top left corner). The scanned slope is located at 30.84° N and 34.75°.
5. Resilience of Zin Valley badlands to climate change The results of erosion tests, regolith mapping and sampling and the DEM analysis reviewed above show that a link between current surface processes, form and landscape development in the Zin Valley badlands is plausible. The data also reconcile the high erosion rates observed on (some) slope sections with the age of the badlands. The key to the improved understanding of the geomorphology of the badlands is a holistic perspective that more than 40 years of research offer to generate an understanding of the spatial patterns of surface characteristic, geomorphic processes, their current rates and
the role of rainfall for runoff and erosion. The spatial differentiation of erosion and deposition processes within a watershed of uniform lithology also questions the close relationship between lithology and surface processes assumed for badlands. Surface erosion by water and mass wasting of slopes are the two basic modes for formation of badlands (e.g. Campbell, 1997; Howard, 1994; Wainwright and Brazier, 2011). The prevalence of either process domain is generally attributed to differences in lithology, most notably water erosion for sandstones and mass wasting for shales. The occurrence of both forms of erosion in the Zin Valley badlands questions this simple conceptual model. The prevalence of runoff discontinuity on large slope sections for
Fig. 13. Regional scale DEM of the western portion of the Zin Valley Badlands drained by the Navarim contributory. Black lines mark the first order catchments and drainage lines.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
A. Yair et al. / Catena xxx (2012) xxx–xxx
Fig. 14. Elongation ratio analysis (Re) extracted from the 5 m regional DEM of the western section of the Zin Valley Badlands.
current rainfall conditions, and the corresponding differentiation of surface characteristics, highlights the role of rainfall as a controlling factor for badland form and process. Consequently, rainfall characteristics appear to control current landscape development in the Zin Valley Badlands. However, this control is not linear, determining only the rate of erosion, as suggested by the model landscape approach, but strongly influenced by the geomorphic history of the badlands. In particular, the formation of colluvial deposits acting as a runoff sink fosters an “autostabilisation” of the slopes, increasing their resilience to climate change. The postulated positive feedback between discontinuous runoff, formation of a colluvium and slope infiltration capacity sheds a critical light on the assumption that up to 300 mm of annual rainfall were required to incise the badlands, limiting their formation to more humid periods (Yair et al., 1980a, 1980b), common in the Levante during the Pleistocene (Issar, 2001). On shorter, juvenile slopes without a colluvium, runoff continuity would be improved, even under current rainfall characteristics (Kuhn, 2011), providing a mechanism for the evacuation of the eroded sediment. In turn, the presence of the colluvium today requires a much greater change in rainfall magnitude to overcome the colluvial runoff sink and increase slope erosion than during the late Pleistocene when incision started on no colluvium had formed. In addition, the frequency of individual rainfall events with a magnitude sufficient for continuous runoff, erosion and sediment export out of first order basins is more important than the annual rainfall amount.
9
The results presented above reconcile the contradiction between high local erosion rates and the age of the badlands by showing that large parts experience little change and are also resilient to change. Thereby, they confirm that the Zin Valley Badlands are indeed not simple landscapes with immediate reactions to climate change. The patterns of the surface characteristics determining infiltration, runoff, erosion and deposition also offer an explanation for the variability of erosional responses observed in other badlands. Basically, one can speculate that without an understanding of the identified scalerelated issues most observations do not deliver data that can be compared between badlands. To do so, slope sections and catchments with similar relevant process domains would have to be observed, rather than plots of a particular size on slope sections with many erosion features. The results discussed above clearly indicate that badlands are more complex than the classic geomorphic literature suggests. However, we would argue that badlands can still be considered as model landscapes. Our main arguments are that despite the spatial heterogeneity of surface properties and processes, badlands are still “simpler” than other landscapes where vegetation and land use modify surface processes. In particular, slopes and fluvial valleys altered by humans represent geomorphic systems in disequilibrium with the driving erosive forces and may thus be rendered particularly sensitive to climate change and high magnitude events. The patchiness of erosion and deposition in badlands also offers conceptual guidance for further questions to be addressed by geomorphologists. For example, the development of a colluvium would lead to a runoff sink due to the increased infiltration capacity of the deposited material compared to less porous and permeable bedrock in any climate. As a consequence, the response to climate change of such a system is muted unless the colluvial buffer is saturated with water or eroded away. So while badlands are more complex than perceived, most landscapes are probably also more complex than we think. Acknowledgments Our thanks go to three generations of researchers working in the Zin Valley badlands. For the most recent work, we would like to acknowledge the support of James Grapes of the Isotope Laboratory at the University of Exeter and the Freie Akademische Gesellschaft for sponsoring the field visit of Nikolaus J. Kuhn and Wolfgang Schwanghart in 2008. References
6. Conclusions The review and results presented above show that the resilience of the Zin Valley Badlands to climate change is relative, depending on the spatial and temporal scale applied. Runoff generation on the ridges is very sensitive to changes in rainfall magnitude, more or less producing a linear positive response to an increase in effective rainfall above the low threshold of runoff generation. The colluvial runoff sink that has formed as a consequence of runoff discontinuity, on the other hand, buffers runoff contribution from slopes to first order channels. As a consequence, a non-linear reaction of slope runoff is likely in response to increasing rainfall. Initially, the frequency of continuous runoff events would increase with more rainfall events above the critical threshold for runoff generation. One can speculate that this would also lead to (i) less deposition and over time even an erosion of the colluvium and (ii) an increase of the extent of rapidly sealing thin crust currently present only at the ridges. These changes in surface characteristics associated with the removal of the colluvial runoff sink would lower the runoff effective rainfall required for continuous runoff from slopes, increasing main channel runoff, sediment export and eventually incision and thus, at least down to the local base level of erosion, rejuvenate the badland formation.
Arkin, Y., Nathan, Y., Starinsky, A., 1972. Paleocene–Early Eocene environments of deposition in the northern Negev. Bulletin — Geological Survey of Israel 56, 1–17. Bentor, Y.K., 1966. The Clays of Israel. Guidebook to the Excursions. The International Clay Conference, Jerusalem, Israel, p. 123. Branson, F.A., Gifford, G.F., Renard, K.G., Hadley, R.F., 1981. Rangeland Hydrology, Range Science Series 1. Society for Range Management, Denver, CO. Bryan, R.B., Yair, A., 1982. Perspectives of studies of badland geomorphology. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich,CT, pp. 1–12. Campbell, I.A., 1997. Badlands and badland gullies. In: Thomas, D.S.G. (Ed.), Arid Zone Geomorphology: Process, Form and Change in Drylands. Belhaven Press, London, Halsted Press, New York, pp. 261–291. Della, Seta M., Del Monte, M., Fredi, P., Palmieri, E.L., 2007. Direct and indirect evaluation of denudation rates in Central Italy. Catena 71, 21–30. Fox, N.I., 2004. Technical Note: the representation of rainfall drop-size distribution and kinetic energy. Hydrology and Earth System Sciences 8 (5), 1001–1007. Gallart, F., Llorens, P., Latron, J., Regüés, D., 2002. Hydrological processes and their seasonal controls in a small Mediterranean mountain catchment in the Pyrenees. Hydrology and Earth Systems Sciences 6 (3), 527–537. Helalia, A.M., Letey, J., Graham, R.C., 1988. Crust formation and clay migration effects in infiltration rate. Soil Science Society of America Journal 52, 251–255. Hodges, W.K., Bryan, R.B., 1982. The influence of material behaviour on runoff initiation in the Dinosaur Badlands, Canada. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich, CT, pp. 13–47. Howard, A.D., 1994. Badlands. In: Abrahams, A.D., Parsons, J.P. (Eds.), Geomorphology of Desert Environments. Chapman & Hall, London, pp. 213–242. Hutchinson, M.F., 1989. A new procedure for gridding elevation and stream line data with automatic removal of spurious pits. Journal of Hydrology 1989 (106), 211–232.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006
10
A. Yair et al. / Catena xxx (2012) xxx–xxx
Issar, Arie S., 2001. Paleo-environments of the uppermost Quaternary in Israel. Current Research, GSI 12, 235–238. Jenny, H., 1941. The Factors of Soil Formation. McGraw-Hill, New York. Kinnell, P.I.A., 2009. The influence of raindrop induced saltation on particle size distributions in sediment discharged by rain-impacted flow on planar surfaces. Catena 78, 2–11. Kuhn, N.J., 2011. Rainfall, runoff and environmental change in the critical zone of two dryland environments. GEOÖKO XXXII, 21–40. Kuhn, N.J., Yair, A., 2003. Spatial distribution of surface conditions and runoff generation in small arid watersheds, Zin Valley Badlands, Israel. Geomorphology 57, 183–200. Kuhn, N.J., Yair, A., Kasanin-Grubin, M., 2004. Spatial distribution of surface properties, runoff generation and landscape development in the Zin Valley Badlands, northern Negev, Israel. Earth Surface Processes and Landforms 29. Kutiel, H., 1978. The distribution of rain intensities in Israel. MSc Thesis, The Hebrew University, Jerusalem. Marks, A.E., 1977. A preliminary overview of central Negev Prehistory, Israel, vol. II. SMU Press, Dallas, pp. 3–31. Morin, J., Cluff, B.C., Powers, W.R., 1970. Realistic rainfall simulation for field investigations. Paper H78. 51st Annual Meeting, American Geophysical Union, Washington, DC. Nadal-Romero, E., Martínez-Murillo, J.F., Vanmaercke, M., Poesen, J., 2011. Scale dependency of sediment yield from badland areas in Mediterranean environments. Progress in Physical Geography June 2011, vol. 35 no. 3, pp. 297–332. Nathan, Y., 1978. Studies on Palygorskite. PH.D.thesis, The Hebrew University of Jerusalem pp 156 (in Hebrew) . Parsons, A.J., Foster, I.D.L., 2011. What can we learn about soil erosion from the use of 137Cs? Earth-Science Reviews 108, 101–113.
Quine, T.A., Walling, D.E., 1993. Use of Caesium-137 measurements to investigate relationships between erosion rates and topography. In: Thomas, D.S.G., Allison, R.J. (Eds.), Landscape Sensitivity. : BGRG Symposia Series. John Wiley and Sons, Chichester, pp. 31–48. Schumm, S.A., 1956. The role of creep and rainwash on the retreat of badland slopes. American Journal of Soil Science 254, 693–706. Schwanghart, W., Kuhn, N.J., 2010. TopoToolbox: a set of Matlab functions for topographic analysis. Environmental Modelling & Software 25, 770–781. Torri, D., Colica, A., Rockwell, D., 1994. Preliminary study of the erosion mechanisms in a biancana badland (Tuscany, Italy). Catena 23, 281–294. Torri, D., Calzolari, C., Rodolfi, G., 2000. Badlands in changing environments: an introduction. Catena 40, 119–125. Wainwright, J., Brazier, R.E., 2011. Slope systems. In: Thomas, D.S.G. (Ed.), Arid Zone Geomorphology: Process, Form and Change in Drylands. John Wiley and Sons, Ltd. Yair, A., Kossovsky, A., 2002. Climate and surface properties: hydrological response of small arid and semi-arid watersheds. Geomorphology 42, 43–57. Yair, A., Lavee, H., Bryan, R.B., Adar, E., 1980a. Runoff and erosion process and rates in the Zin Valley badlands, Northern Negev, Israel. Earth Surface Processes 5, 205–225. Yair, A., Lavee, H., Goldberg, P., Bryan, R.B., 1980b. Present and past geomorphic evidences in the development of a badlands landscapes, Zin valley Badlands, northern. Yair, A., Goldberg, P., Brimer, B., 1982. Long term denudation rates in the Zin–Havarim badlands, northern Negev, Israel. In: Bryan, R.B., Yair, A. (Eds.), Badland Geomorphology and Piping. Geobooks, Norwich, CT, pp. 279–292. Zapata, F. (Ed.), 2002. Handbook for the assessment of soil erosion and sedimentation using environmental radionuclides. : The chapter in question is Chapter 4. Kluwer Acdemic Publishers, London, pp. 59–65.
Please cite this article as: Yair, A., et al., The resilience of a badland area to climate change in an arid environment, Catena (2012), doi:10.1016/j.catena.2012.04.006