Runoff-induced sediment yield over dune slopes in the Negev Desert. 2: Texture, carbonate and organic matter

Chapter 17Runoff and Erosion Processes Within a Dune System

G.J. Kidron and A. Yair(*ü )

17.1 Introduction

The occurrence of surface runoff processes is often regarded by geomorphologists or hydrologists as irrelevant within sandy areas and dune systems, due to the very high infiltration rates of sand. The presumed lack of runoff processes has been mentioned several times in the case of the northwestern Negev sand field (Hillel and Tadmor 1962; Tsoar and Zohar 1985; Tsoar and Møller 1986). The high infil-tration rates of over 100 mm h−1 for loose sand, far above the common range of rain intensities in the Negev, may explain the assumed unlikelihood of runoff genera-tion within the Negev dune systems. However, field observations have drawn atten-tion to the existence of widespread biological topsoil crusts (Danin 1978; Tsoar 1990; Yair 1990), rich in fine-grained particles that are expected to reduce infiltra-tion rate (Chaps. 10 and 18, this volume).

In order to obtain an initial insight into the issue of runoff generation in the Nizzana sandy area, a sprinkling experiment, with an intensity of 18.4 mm h−1, was conducted over a small plot covering 1.5 m2 (Fig. 17.1). The experiment was per-formed in the winter time, under wet surface conditions. Runoff developed within 3 minutes, after only 1 mm of rain. Final infiltration rate was reached quickly, and amounted at ~12 mm h−1 (Yair 1990). Shortly after this run, a second sprinkling experiment was performed over the same plot after the topsoil crust had been removed. Rain intensity in the second run was increased to 53 mm h−1. Despite the antecedent wet conditions, and the extremely high rain intensity applied, no runoff was observed for 42 minutes, by which time the accumulated rain amount reached the value of 37 mm.

The results of the two experiments clearly demonstrated the importance of the biological crust in limiting infiltration and enhancing runoff generation. It was therefore postulated that the microbiotic crust may affect the hydrological behavior of the surface under natural rain conditions, and the spatial distribution of water resources in the study area.

The results of the sprinkling experiments cannot adequately represent processes under natural rainfall conditions, for the following reasons:

S.-W. Breckle et al. (eds.), Arid Dune Ecosystems. Ecological Studies 200. 239© Springer-Verlag Berlin Heidelberg 2008

Breckle_Ch17.indd 239Breckle_Ch17.indd 239 11/29/2007 9:56:21 AM11/29/2007 9:56:21 AM

240 G.J. Kidron, A. Yair

1. The sprinkled area covers only 1.5 m2, where the topsoil crust is relatively uni-form, while under field conditions crust thickness and composition may vary along the dune slope as well as laterally.

2. The sprinkling was conducted under wet surface conditions, making it impossi-ble to verify whether the crust possesses water-repellent properties, usually very pronounced under dry conditions (Rutin 1983). Water repellency would be expected to limit infiltration rate, and enhance surface runoff generation.

3. Finally, rainfall was applied at a uniform intensity, while natural rainstorms in the area are intermittent, and characterized by high temporal variability in rain intensities.

17.2 Field Instrumentation and Methodology

For the reasons listed above, the research site was equipped with devices for the measurement of rainfall and runoff. Runoff plots were constructed on north (N)- and south (S)-facing slopes in the upper, middle, and lower parts of the dune slopes (Fig. 17.2). The location of the plots, and their characteristics are given in Fig. 17.3 and Table 17.1. Runoff was measured either with a V-notch weir and a pressure gauge transducer or by collection in large containers. Rainfall was recorded with an electronic tipping-bucket recorder. Ten large plots were constructed, using 0.5-mm-thick metal sheets, 20 cm high, inserted 10 cm into the sand. Four of the large plots were constructed on the north-facing slope (N1–N4) and four on the south-facing slope (S1–S4). Two additional large plots (C1, C4) were constructed on the north-facing slope of a low, stabilized dune. Of the four plots constructed on each of the active (i.e., having a mobile crest) dunes, two plots were subdivided into three subplots: N1.1 and N2.1 included the upper active section of the north-facing

Fig. 17.1 Results of sprinkling experiment


17 Runoff and Erosion Processes Within a Dune System 241

Fig. 17.2 View of runoff plots

Fig. 17.3 Layout of the experimental site

exposure, while S1.1 and S3.1 drained the upper active area of the south-facing exposure. Likewise, N1.2 and N2.2 on the northern exposure, and S1.2 and S3.2 on the southern exposure drained the semi-stabilized mid-slopes. N1.3, N2.3, and S1.3, S3.3 drained the lower flanks of the northern and southern exposure, respec-tively; stabilized by a continuous topsoil biological crust. Strips 0.6–1.2 m wide



separated the subunits to enable plot construction and passage without destroying the crusts within the plots. All other plots drained the entire slope length.

Four plots and subplots (N2.3, N3, S2, and S3.3) were equipped with V-notch weirs, and pressure transducers connected to data loggers. All other plots were equipped with large containers. Where large runoff volumes were expected, a split-ting device was used. The splitting device enabled the capture of 10% runoff water and suspended sediment. Runoff and sediment collected are considered representa-tive of runoff water and sediment allowed to flow away.

Rainfall was measured using an electronic tipping-bucket rain recorder (Texas Electronics, USA; accuracy 0.1 mm), connected to a CR-10 Campbell data logger. Runoff and sediment were measured following each rainstorm. Sediments collected were oven-dried at 105 °C until reaching a constant weight, and then weighed. A representative sample of approximately 30 g was wet-sieved with 0.5% sodium hexametaphosphate (to ensure clay separation) through a 62-µm mesh, and the amounts of sand, silt, and clay were determined.

Rainfall, runoff, and sediment yield measurements were carried out during 1990–1994. Owing to differences in plot dimensions, as well as to differences in surface characteristics along the slope, the unit chosen herein for the comparison of runoff from large plots is based on runoff volume per plot width. Likewise, sedi-ment yield will be also presented per plot width, following Rutin (1983). This presentation enables comparison without any a priori knowledge of the actual plot length, and area involved in runoff and sediment contribution. Needless to say, dif-ferences in surface properties are more pronounced in large than in medium-sized and small plots.

17.3 Results

Annual precipitation exhibits a high variability (46.9–131.4 mm), with a limited number of rainstorms (10–20). Most storms were small, over 60% yielding less than 5 mm (Fig. 17.4A). High rain intensities of over 30 mm h−1 lasted for up to 8 minutes.

Table 17.1 Main characteristics of runoff plots

Plot number Plot characteristics Plot location

Subdivided large plots N1.1, N2.1, S1.1, S3.1 No crust, < 5% vegetation

Dune crest, active dune

N1.2, N2.2, S1.2, S3.2 40–75% patchy crust, 10–30% vegetation

Mid-dune, active dune

N1.3, N2.3, S1.3, S3.3 > 95% crust, 10–30% vegetation

Dune bottom, active dune

Whole-slope plots N3, N4, S2, S4 0–97% crust, 10–40% vegetation

Entire slope, active dune

C1,C4 > 95% crust, 10–20% vegetation

Entire slope, stabilized dune



The maximum rain intensity recorded was 72 mm h−1, lasting for 1 minute only. Medium and high rain intensities of ≥ 9 mm h−1 accounted for approximately 20% of the precipitation (Fig. 17.4B). The storms were characterized by intermittent rain spells of variable intensities (Fig. 17.5).

Runoff generation was intermittent, with 1–8 independent flows during a single rainstorm. Whereas medium and high rain intensities were capable of runoff gener-ation, low rain intensities of < 9 mm h−1 were not (Fig. 17.5). Nevertheless, not all medium- and high-intensity rain spells resulted in runoff. Runoff was not generated during the onset of most rainstorms, during which the infiltration capacity of the dry crusts usually exceeded medium and high rain intensities. High infiltration rates under dry surface conditions are indicative of the fact that the biological topsoil crusts in the study area are not water-repellent. Rather, runoff generation resulted from pore clogging (Kidron et al. 1999). Infiltration was impeded as a result of sheaths and slime swelling following water absorption (Verrecchia et al. 1995;

Fig. 17.4 Frequency distributions of rainstorm depths (A) and rain intensities (B) during the period 1989–1994



Mazor et al. 1996; Kidron and Yair 1997). Infiltration, and consequently runoff were not uniform throughout the area, and showed high spatial and temporal varia-bility (Chap. 18, this volume). Runoff, and consequently sediment flow were not generated at the mobile crest devoid of crust. Runoff and sediment flow were, however, generated at the mid- and foot-slope plots.

Runoff volumes and sediment yields recorded at the mid- and foot-slope subplots of north- and south-facing plots are shown in Fig. 17.6. In both cases, runoff and sedi-ment yield were significantly higher at the foot-slope than at the mid-slope plots (paired t-test; p < 0.05). The absence of runoff from the mobile dune sections, the low amounts obtained from the mid-slopes, and the much higher amounts obtained from the foot-slope plots clearly indicate that runoff yield, and consequently sediment yield are controlled by the crust cover (Yair 1990; Kidron and Yair 1997; Yair 2001).

The runoff volumes recorded were significantly higher at north-facing than at south-facing plots (Fig. 17.7). On average, the north-facing semi-active and stabi-lized dune sections yielded 16.1 and 15.7 l m−1, respectively, compared to 5.0 l m−1 for the south-facing slope (Kidron 1999). This is due to the better development and spatial continuity of the biological crust at the former than at the latter plots. Plot N1.3 generated the highest runoff, explained by its very smooth surface, and rela-tively low vegetation cover.

For sediment yields, the trends are similar to those recorded for runoff. The highest amounts of sediments were obtained at the north-facing plots. This result may be explained by the higher capacity of runoff to carry sediment. Plot N3 often yielded the highest values (Fig. 17.8). This plot drains a whole plot in which the upper part is characterized by a blowout. Loose sand particles, blown by the prevalent southwesterly winter winds, were deposited over the crusted area, and washed out by runoff into the sediment collector. Average annual sediment yield

Fig. 17.5 Rainstorm of 1–3 January 1992



Fig. 17.6 Average annual runoff volumes at mid- and foot-slope plots of north- and south-facing slopes

Fig. 17.7 Average annual runoff volumes for all plots (1990–1994)

on the active and stabilized north-facing slopes were 795 and 431 g m−1, respec-tively, compared to 82 g m−1 on the south-facing slopes (Kidron and Yair 2001). This process, combined with the limited runoff characteristic of the mid-slope sections, explains why higher sediment concentrations were obtained at the mid-slope



than at the foot-slope plots (Fig. 17.9). Sediment concentrations obtained in the mid-slope section sometimes reached the value of 500 g l−1. This is due to the fact that a higher proportion of the sediment collected at the mid-slope plots is com-posed of heavy sand particles; while sediment collected at the foot-slope plots is richer in fines (up to 30%) derived from the topsoil biological crust, rich in fine-grained particles.

17.4 Discussion

The notion that the development of arid sand dunes is solely a result of eolian activ-ity is widespread in the literature. Many researchers have assumed that runoff does not take place in arid sand dunes, and consequently no erosion or sedimentation by overland flow can be expected. However, microbiotic crusts are widespread in many arid and semiarid parts of the world (see West 1990, and Belnap and Lange 2001 for reviews). Once present, the crusts may significantly alter the hydrological behavior of the surface. Data obtained in the Nizzana area show that runoff yield, and consequently sediment yield are positively linked to crust cover and biomass,

Fig. 17.8 Relationships between sediment yield and plot area



Fig. 17.9 Average sediment concentrations at a north- and b south-facing plots

which in turn are linked to exposure and topography (Yair 1990; Kidron and Yair 1997, 2001). Whereas the crest of the active dunes lacked microbiotic crusts, and as a result did not generate runoff, intermediate runoff and sediment yields charac-terized the mid-slope sections having a thin and patchy crust cover. Furthermore, differences in crust biomass and spatial continuity are responsible for the differential behavior of north- and south-facing slopes. Runoff and sediment yields are higher on the former. North-facing crusts are indeed characterized by 2 to 3 times higher chlorophyll a contents than are south-facing crusts (Kidron 1995). Higher crust biomass, characterized by a dense network of sheaths and filaments, promotes pore clogging responsible for higher runoff yield on the bottom north-facing slopes. The larger runoff amounts, and the supply of loose sand by the prevailing southwesterly winds may also explain the higher average annual sediment yield on the north-facing slopes, this being on average one order of magnitude higher than that of the south-facing plots.

Owing to the abrupt change in slope angle in the interface dune–interdune section, most runoff, and hence sediments will concentrate along a narrow (usually



3–5 m) belt in the north-facing foot-slope area. In many cases, runoff and sediments will be concentrated within small local depressions usually 1–4 m large, charac-terized by a dense vegetation cover. The dune–interdune interface along the north-facing foot slopes can be regarded as the most mesic habitat within the dune field. It is inhabited by a moss-dominated crust. This crust has a chlorophyll a content of 50–60 mg m−2, compared to only 30–40 mg m−2 for the other north-facing slopes, and to only 15–20 mg m−2 for the south-facing slopes and interdune corridors. It also has the highest variety of cyanobacteria and green algae (Kidron et al. 2000).

17.5 Conclusions

The notion that runoff generation, and hence runoff-transported sediments do not occur in arid dune fields is not supported at the Hallamish dune field. The complex interrelationships between abiotic and biotic factors, and between eolian and fluvial factors that contribute to the development of the crust, and to its role within the Hallamish dune-field ecosystem can be described as follows:

1. Erosion and deposition by wind may form and alter the topography, creating longitudinal dunes and blowouts. The wind may also affect the establishment and cover of microbiotic crusts, preventing crust establishment in places sub-jected to high erosion or deposition, and facilitating crust establishment in areas having a relatively low eolian activity. Wind activity, topography, and slope aspect will affect the length of time during which the surface is wet, and consequently will affect crust development, infiltration rates, and runoff generation.

2. Once established, the crust may impede wind erosion. The crust may also alter the hydrological properties of the surface, affecting water redistribution and wetness duration.

3. Runoff will in turn affect crust and plant growth, as well as sediment production. Sediment production will also be influenced by topography, i.e., slope angle, and the presence of blowouts.

4. Erosion and sedimentation by runoff will alter the topography, resulting in local sedimentation along the foot slopes. Although runoff and erosion may be very low in this case, compared to that of sand dunes in wet climates (Rutin 1983), or of other types of surfaces in the Negev desert (Yair 1974), their role in shaping the dune-field ecosystem cannot not be ignored. Both runoff and sediment yield will affect plant density and biomass, as well as animal distribution.

Acknowledgements The research was supported by the Arid Ecosystem Research Centre (AERC) of the Hebrew University of Jerusalem and the MINERVA foundation, and by a grant from DISUM. We would like to thank E. Sachs for his assistance in the field, M. Kidron for the drawings, and C.A. Kidron for editing the manuscript.



References

Belnap J, Lange OL (eds) (2001) Biological soil crusts: structure, function, and management. Ecological Studies vol 150. Springer, Berlin Heidelberg New York

Danin A (1978) Plant species diversity and plant succession in a sandy area in the Northern Negev. Flora 167:409–422

Hillel D, Tadmor N (1962) Water regime and vegetation in the Central Negev highlands of Israel. Ecology 43:33–41

Kidron GJ (1995) The impact of microbial crust upon rainfall-runoff-sediment yield relationships on longitudinal dune slopes, Nizzana, western Negev Desert, Israel (in Hebrew with English summary). PhD Thesis, The Hebrew University of Jerusalem

Kidron GJ (1999) Differential water distribution over dune slopes as affected by slope position and microbiotic crust, Negev Desert, Israel. Hydrol Processes 13:1665–1682

Kidron GJ, Yair A (1997) Rainfall-runoff relationships over encrusted dune surfaces, Nizzana, western Negev, Israel. Earth Surface Processes Landforms 22:1169–1184

Kidron GJ, Yair A (2001) Runoff-induced sediment yield over dune slopes in the Negev Desert. 1. Quantity and variability. Earth Surface Processes Landforms 26:461–474

Kidron GJ, Yaalon DH, Vonshak A (1999) Two causes for runoff initiation on microbiotic crusts: hydrophobicity and pore clogging. Soil Sci 164:18–27

Kidron GJ, Barzilay E, Sachs E (2000) Microclimate control upon sand microbiotic crust, western Negev Desert, Israel. Geomorphology 36:1–18

Mazor G, Kidron GJ, Vonshak A, Abeliovich A (1996) The role of cyanobacterial exopolysac-charides in structuring desert microbial crusts. FEMS Microbiol Ecol 21:121–130

Rutin J (1983) Erosional processes on a coastal sand dune, De Blink, Noordwijkerhout. Physical Geography and Soils Laboratory, University of Amsterdam, Publ no 35, pp 1–144

Tsoar H (1990) The ecological background, deterioration and reclamation of desert dune sand. Agric Ecosystem Environ 33:147–170

Tsoar H, Møller JT (1986) The role of vegetation in the formation of linear sand dunes. In: Nickling WG (ed) Aeolian geomorphology. Proc 17th Annual Binghamton Geomorphology Symp, Allen and Unwin, Boston, MA, pp 75–95

Tsoar H, Zohar Y (1985) Desert dune sand and its potential for modern agricultural development. In: Gradus Y (ed) Desert development. Reidel, Boston, MA, pp 184–200

Verrecchia E, Yair A, Kidron GJ, Verrecchia K (1995) Physical properties of the psammophile cryptogamic crust and their consequences to the water balance of sandy soils, Northwestern Negev Desert, Israel. J Arid Environ 29:427–437

West NE (1990) Structure and function of microphytic soil crusts in wildland ecosystems of arid to semi-arid regions. Adv Ecol Res 20:180–223

Yair A (1974) Sources of runoff and sediment supplied by the slopes of a first order drainage basin in an arid environment. Report on present day geomorphological processes. Abhandlungen der Akademie der Wissenshaften Göttingen, Mathematische-Physikalische Klasse no 29, pp 403–407

Yair A (1990) Runoff generation in a sandy area – the Nizzana sands, Western Negev, Israel. Earth Surface Processes Landforms 15:597–609

Yair A (2001) Effects of biological soil crusts on water redistribution in the Negev desert, Israel. A case study in longitudinal dunes. In: Belnap J, Lange OL (eds) Biological soil crusts: struc-ture, function, and management. Ecological Studies vol 150. Springer, Berlin Heidelberg New York, pp 303–314



Runoff-induced sediment yield over dune slopes in the Negev Desert. 2: Texture, carbonate and organic matter

Documents