Negev Rockfalls

Maps, Aerial Views, Illustrations, Charts, and Photos

Timna Rockfalls

Maps

Normal Size

Fig. 1a Location Map

Fig. 1a

DEM of Dead Sea fault (DSF) region showing the location of the study area. The DSF system is located between the parallel dashed lines.

Matmon et al. (2005)

showing study area in relation to the Dead Sea fault (DSF) region from Matmon et al. (2005)
Fig. 1b Aerial photo of the study area
Fig. 1b

Aerial photo of the study area.

JW:
- E = 3rd Site
- C = Lower Valley
- D = Upper Valley
Matmon et al. (2005)
of the study area from Matmon et al. (2005)

Magnified

Fig. 1a Location Map

Fig. 1a

DEM of Dead Sea fault (DSF) region showing the location of the study area. The DSF system is located between the parallel dashed lines.

Matmon et al. (2005)

showing study area in relation to the Dead Sea fault (DSF) region from Matmon et al. (2005)
Fig. 1b Aerial photo of the study area

Fig. 1b

Aerial photo of the study area.

Matmon et al. (2005)

of the study area from Matmon et al. (2005)

Aerial Views

Timna "Upper Valley"

Timna "Upper Valley" Rockfalls Area

click on image to explore this site on a new tab in Google Earth

Rockfalls Area in Google Earth
Timna "Lower Valley"

Timna "Lower Valley" Rockfalls Area

click on image to explore this site on a new tab in Google Earth

Rockfalls Area in Google Earth
Timna "Upper Valley"

Timna "Upper Valley" Rockfalls Area

click on image to explore this site on a new tab in govmap.gov.il

Rockfalls Area on govmap.gov.il
Timna "Lower Valley"

Timna "Lower Valley" Rockfalls Area

click on image to explore this site on a new tab in govmap.gov.il

Rockfalls Area on govmap.gov.il

Illustrations and Photos

Lower Valley Site - Timna-1, Timna-2, and Timna-3

Sample Locations

from Matmon et al. (2005)

Fig. 1c

Lower Valley site. Photo taken looking towards the east. Source cliff is on the left. Person on left below boulder for scale. Cosmogenic age uncertainties include analytical errors, a 10% uncertainty in production rate, and sample-to-sample variation.

Matmon et al. (2005)

Cliff and Boulder Faces at Timna-1

from Matmon et al. (2005)

Fig. 2

Close-up of the detailed structure of the cliff and boulder faces. Small (1–5 cm) quartzite pebbles on the boulder face can be matched to cavities on the cliff face and vice versa. This perfect match enables identification of the exact location of the boulder before it was detached. The perfect match also indicates insignificant erosion since detachment. Rock pick for scale. Several cavities can be seen below the pencil.

Photo of the boulder from which Timna 1 (in the center) was sampled. The relief of fracture ribs is apparent and indicates insignificant erosion since detachment of the boulder. Person for scale

Matmon et al. (2005)

Sample Locations Illustrations

from Matmon et al. (2005)

Fig. 3a

Cartoon showing the sequence of events that formed the present day morphology of the Lower Valley site.

Initial valley before boulder formation
After the formation of the first boulder pile, ~15 ka, the small valley was only partially blocked and sediment did not accumulate for several thousands of years
The initial deposition of sediment, ~6.5 ka, is recorded by the lower unit in pit 1 upstream of the first pile of boulders
The deposition of the "Closer Pile" at ~4 ka and the deposition of the second and third sedimentary units (recorded in the upper unit in pit 1 and in both units in pit 2)

Thin dashed line — the original position of the valley wall
Thick dashed line — the position of the valley wall after the deposition of the "Farther Pile"

Matmon et al. (2005)

Upper Valley Site - Timna-10 and Timna-11

Sample Locations

from Matmon et al. (2005)

Fig. 1d

Upper Valley site. Photo taken looking towards the south. People on boulder for scale. Cosmogenic age uncertainties include analytical errors, a 10% uncertainty in production rate, and sample-to-sample variation.

Matmon et al. (2005)

Photos and Illustration

from Matmon et al. (2005)

Fig. 4

Boulder pile at the "Upper Valley" location. Dated boulder is marked by dashed lines. Bedrock block (5 m x 1.5 m x 2 m; marked by the letter A) is bounded by fractures.
The back fracture of block A is open ~30 cm and suggests the movement of the boulder in the direction of the cliff’s free face.
Cross-section through the bedrock block perpendicular to the cliff face shows the gentle angle of the base fracture of the block. This angle (LT 10°) indicates that movement requires ground acceleration

Matmon et al. (2005)

Third Site - TMN-7, TMN-8, and TMN-9

Sample Locations

from Matmon et al. (2005)

Fig. 1e

Third site location. Single boulders at increasing distance from source cliff. The degree of weathering increases with distance from the cliff. The age of boulder TMN-9 was not determined, however, its distance from the cliff (4 m) helps estimate the average rate of retreat of the cliff. Person near boulder TMN-8 for scale. Cosmogenic age uncertainties include analytical errors, a 10% uncertainty in production rate, and sample-to-sample variation.

Matmon et al. (2005)

Age Chart for All Timna Rockfall Sites

from Matmon et al. (2005)

Fig. 3b

Age relations between cosmogenic nuclide and OSL samples. The correlation between the OSL samples is not certain apart from the significant difference between the lower sedimentary unit in pit 1 and the upper sedimentary unit in pit 2. In the "Closer Pile", only sample Timna 3 can be statistically old enough to be older than the sediments and represent the time of boulder detachment. The dashed horizontal lines indicate the most probable ages of events that could have caused the detachment of the boulders sampled by Timna 1, Timna 2, Timna 10, and TMN-8.

Matmon et al. (2005)

Shehoret Rockfalls

Maps

Normal Size

Fig. 1 Location Map

Fig. 1

General location of the research area in the eastern Mediterranean (black star) and an aerial orthophoto of the study sites in the vicinity of Nahal Shehoret. All sites are located within a few 100s of meters of each other.

Rinat et al. (2014)

for Rockfall sites SH1, SH2, SH4, and SH10 in the vicinity of Nahal Shehoret from Rinat et al. (2014)

Magnified

Fig. 1 Location Map

Fig. 1

General location of the research area in the eastern Mediterranean (black star) and an aerial orthophoto of the study sites in the vicinity of Nahal Shehoret. All sites are located within a few 100s of meters of each other.

Rinat et al. (2014)

for Rockfall sites SH1, SH2, SH4, and SH10 in the vicinity of Nahal Shehoret from Rinat et al. (2014)

Illustrations and Photos

Site SH1

Sample Locations

from Rinat et al. (2014)

Left Pre-fall Setting (Fig. 5)

Crosses represent sample locations
Stars represent samples projected to the cross section plain
Pre-fall depths and distance from cliff wall can be seen in the figure

Middle OSL samples (Fig. 2b)

OSL samples, marked with black dots, were collected from post-fall sediments that abut the boulders or accumulated underneath boulders.

Right TCN samples (Fig. 2a)

TCN samples were collected from the bedrock at the upper surface above each rockfall site (samples RF), from the rockfall scar on each cliff (samples BR), and from exposed boulder faces. Paired samples in each site are marked with identical polygons.

Rinat et al. (2014)

Site SH2

Sample Locations

from Rinat et al. (2014)

Left Pre-fall Setting (Fig. 5)

Crosses represent sample locations
Stars represent samples projected to the cross section plain
Pre-fall depths and distance from cliff wall can be seen in the figure

Middle OSL samples (Fig. 2b)

OSL samples, marked with black dots, were collected from post-fall sediments that abut the boulders or accumulated underneath boulders.

Right TCN samples (Fig. 2a)

TCN samples were collected from the bedrock at the upper surface above each rockfall site (samples RF), from the rockfall scar on each cliff (samples BR), and from exposed boulder faces. Paired samples in each site are marked with identical polygons.

Rinat et al. (2014)

Site SH4

Sample Locations

from Rinat et al. (2014)

Left Pre-fall Setting (Fig. 5)

Crosses represent sample locations
Stars represent samples projected to the cross section plain
Pre-fall depths and distance from cliff wall can be seen in the figure

Middle OSL samples (Fig. 2b)

OSL samples, marked with black dots, were collected from post-fall sediments that abut the boulders or accumulated underneath boulders.

Right TCN samples (Fig. 2a)

TCN samples were collected from the bedrock at the upper surface above each rockfall site (samples RF), from the rockfall scar on each cliff (samples BR), and from exposed boulder faces. Paired samples in each site are marked with identical polygons.

Rinat et al. (2014)

Site SH10

Sample Locations

from Rinat et al. (2014)

Left Pre-fall Setting (Fig. 5)

Crosses represent sample locations
Stars represent samples projected to the cross section plain
Pre-fall depths and distance from cliff wall can be seen in the figure

Middle OSL samples (Fig. 2b)

OSL samples, marked with black dots, were collected from post-fall sediments that abut the boulders or accumulated underneath boulders.

Right TCN samples (Fig. 2a)

TCN samples were collected from the bedrock at the upper surface above each rockfall site (samples RF), from the rockfall scar on each cliff (samples BR), and from exposed boulder faces. Paired samples in each site are marked with identical polygons.

Rinat et al. (2014)

Schematic Cross-Section

from Rinat et al. (2014)

Fig. 4

Schematic cross-section across site SH10. Remnants of alluvial sediments (named hereT-1and T-2) suggest earlier coverage and burial of the site by early to middle Pleistocene terraces of Nahal Shehoret. The T-1 and T-2 terraces share similar characteristics such as sediment type, appearance, and roughness. Connecting the two terraces suggests minimum sediment coverage above the site of ~15 m.

Rinat et al. (2014)

Matching Boulders

from Rinat et al. (2014)

Fig. 3

Site SH10 and two of its ~75 m³ matching boulders, located at the foot of their source cliff; note a person (marked by a circle) for scale. The once-connected boulders match perfectly, as seen by matching lithoclasts found on the presently exposed faces (b, c). Marks like these enable the matching of faces on different boulders and on the cliff, ensure sampling from the same pre-detachment locations and provide the necessary evidence for negligible erosion.

Rinat et al. (2014)

Map and Chart of Holocene Seismic Events from Paleoseismic Sites on the DSF

from Rinat et al. (2014)

Fig. 10

Evidences for Holocene seismic events from paleoseismic records along the length of the DSF. The time interval, represented by a gray box, is determined by the maximum age range of the four dated boulder piles in this study. Possible seismic events are recorded at ~5 ka along the entire length of the Dead Sea rift and its main branch the Carmel fault.

Locations of the sites mentioned in (a)

Rinat et al. (2014)

Timna Rockfalls - Seismic Events at ~31 ka and ~15 ka

Discussion

Matmon et al. (2005:811) used the combined exposure ages determined from cosmogenic isotopes and OSL ages from three rockfall sites at Timna to suggest several boulder detachment events which they surmise were caused by earthquakes. The oldest event (31.0±4.9 ka) was recorded only in one of the three sites (named the third site). The next oldest event apparently caused rockfalls in all three Timna sites. A summary of the dates for the next oldest event is summarized in the Table below:

Age	Site	Comments
15.1±3.4 ka	Upper Valley	Samples Timna-10 and Timna-11
15.1±2.1 ka	Lower Valley	Sample Timna-1
13.9±4.8 ka	Third Site	boulder situated 14 meters from the cliff in the third site dated by OSL within 1σ of Timna-1 date from Lower Valley

References

Matmon et al. (2005)

Abstract

from Matmon et al. (2005:803-804)

In this study, we explored the spatial and temporal relations between boulders and their original in-situ locations on sandstone bedrock cliffs. This was accomplished by combining field observations with dating methods using cosmogenic isotopes (¹⁰Be and ¹⁴C ) and optically stimulated luminescence (OSL). Our conclusions bear both on the landscape evolution and cliff retreat process in the hyperarid region of Timna and on the methodology of estimating exposure ages using cosmogenic isotopes.

We recognize three discrete rock fall events, at 31 ka, 15 ka, and 4 ka. In this hyperarid region, the most plausible triggering mechanism for rock fall events is strong ground acceleration caused by earthquakes generated by the nearby Dead Sea fault (DSF). Our record, however, under represents the regional earthquake record implying that ongoing development of detachment cracks prior to the triggering event might be slower than the earthquake cycle.

Cliff retreat rates calculated using the timing of rock fall events and estimated thickness of rock removed in each event range between 0.14 m ky⁻¹ and 2 m ky⁻¹. When only full cycles are considered, we derive a more realistic range of 0.4 m ky⁻¹ to 0.7 m ky⁻¹. These rates are an order of magnitude faster than the calculated rate of surface lowering in the area. We conclude that sandstone cliffs at Timna retreat through episodic rock fall events that preserve the sharp, imposing, landscape characteristic to this region and that ongoing weathering of the cliff faces is minor.

A 10%–20% difference in the ¹⁰Be concentrations in samples from matching boulder and cliff faces that have identical exposure histories and are located only a few meters apart indicates that cosmogenic nuclide production rates are sensitive to shielding and vary spatially over short distances. However, uncertainties associated with age calculations yielded boulder and matching cliff face ages that are similar within 1σ . The use of external constraints in the form of field relations and OSL dating helped to establish each pair's age. The agreement between calculated ¹⁴C and ¹⁰Be ages indicates that the accumulation of ¹⁰Be at depth by the capture of slow deep-penetrating muons was properly accounted for in the study.

Introduction

from Matmon et al. (2005:804-805)

Cliffs and unstable slopes are sources of detritus for a variety of mass-movement processes. Rock falls occur when recently detached segments of bedrock slide down slope. This is likely to occur along steep mountain sides and may have catastrophic results. A detached block falls from a rock wall when the gravity-determined shear stress is greater than the shearing resistance [1]. The detachment fissure separating the block from the outcrop is usually already present.

Tectonic, climatic, environmental, and random factors contribute to the occurrence of rockslides. Many rock falls are caused by undercutting and steepening of the rock wall because of lateral stream erosion or accelerated back weathering of the wall foot. Nevertheless, the triggering of a block fall or a rock fall after an extended period of preparation is often a minor event such as a small earthquake, lightning, or heavy rainfall [1]. The rate at which rock falls occur may determine the rate of cliff retreat and the potential hazard associated with these falls. Thus, dating past rock falls can help to better understand the process of cliff retreat. To estimate the rate of cliff retreat, it is essential to determine the position of the cliff at some time in the past and examine the processes by which it arrived at its present position.

Dating boulders, mainly with cosmogenic radio nuclides (CRN), has become common practice in geomorphologic studies. On active moraines and alluvial fans, deposited boulders might contain CRN from prior exposure and might be buried, re-exposed and rolled. Thus, boulder ages represent the time at which they become stable elements in the landscape [2,3]. The time between deposition and stabilization might be in the order of 10²–10³ years. In the case of rock falls, boulders are deposited instantaneously. Previously shielded rock faces are exposed, and their ages correlate to discrete geomorphic events. In arid environments, where erosion rates are slow, fresh and well-preserved exposed rock faces represent the actual time of the rock fall events. Rock piles may indicate the general past position of a cliff. Thus, we hypothesize that by dating rock piles, or individual boulders in such piles, we open a window that enables us to better understand landscape development in arid environments controlled by rock falls and other cliff retreat processes.

In the present study, we describe the spatial and temporal relations between very large angular blocks (which will be referred to as boulders) and their original in-situ locations on sandstone bedrock cliffs along the western margins of the DSF, southern Israel (Fig. 1). Boulder detachment events were dated by combining exposure age dating using cosmogenic isotopes (¹⁰Be, ¹⁴C) and by OSL. The results indicate that boulder detachment events in this area occur less often than GT 6 M earthquakes [4]. Rock falls occur only when boulders are sufficiently detached from the bedrock cliffs by large and developed cracks and are thus susceptible to the next triggering event.

Geological and morphological setting

from Matmon et al. (2005:805-806)

Timna is located at the western margin of the southern Arava Valley, a hyperarid region extending from the Dead Sea basin to the Gulf of Aqaba (Fig. 1). Timna is an uplifted core of late Precambrian igneous rocks [5] overlain by Cambrian and Lower Cretaceous clastic sediments. It is a bowl-shaped valley, about 8 km in diameter, and is surrounded by cliffs of Upper Cretaceous limestone and dolomite rising up to 600 m above the valley floor.

Landscape development in the Timna area is controlled by the tectonic activity along the southern section of the DSF (Hannan Ginat, personal communication). Erosional truncation of the Upper Cretaceous carbonate cap rocks during the Oligocene and Miocene exposed the more erodable Lower Cretaceous sandstone. As the southern Arava Valley developed into a topographic base level, the sandstone was eroded and transported towards that base level, the Timna Valley was incised, and the Precambrian basement and Cambrian sediments were exposed at the base of the Timna Valley.

The Cambrian sandstone in Timna reaches a thickness of about 100 m [5] and forms cliffs that dominate the landscape. Within the Cambrian sandstone terrain, erosion is dominated both by detachment of massive boulders from the sandstone cliffs and by slow weathering of the cliff faces, as evident from 0.1-m-scale cavities (tafoni). Generally, boulders detached from sandstone cliffs are found in piles and are usually coated with varnish. In many locations, the cliff face and the faces on the boulders match perfectly both in detail and overall morphology. Some of these boulders can be accurately traced to their original position on the cliff from which they were detached (Fig. 2).

We concentrated on three boulder sites in the Timna area (Fig. 1). The first, the Lower Valley, contains two boulder piles (one close to the source sandstone cliff face, which we call the "Closer Pile", and one farther away, the "Farther Pile") detached from a cliff that rises 10–14 m above the valley floor. The boulders at this site were deposited at the mouth of the Lower Valley (30 x 60 m), creating a natural dam, and causing upstream accumulation of sandy sediments on the valley floor. The second site, the Upper Valley, contains several boulder piles. We concentrated on a single boulder pile detached from a nearby 10 m cliff. At both locations boulders are not significantly affected by erosion and perfectly match the cliff faces (Fig. 2). Unlike the first two sites that contain unweathered boulder piles, the third site we examined contains three weathered boulders that are located at distances of 20, 14, and 4 m from the source cliff (Fig. 1E).

Discussion

Boulder ages and boulder detachment events

from Matmon et al. (2005:810-811)

Both in the Upper and Lower Valley sites, a boulder and its matching cliff face were sampled. In both cases, the sampled boulder can be matched perfectly back to its original location on the cliff. The assumed history of the boulders and their matching cliff faces imply an identical exposure age for each pair. After detachment, the boulders differed from the cliff faces in their topographic shielding and geometric properties. Therefore, production rates of cosmogenic nuclides in the boulders were different from Valley, a 10% difference in ¹⁰Be concentration between the boulder and its matching cliff face was measured; a 20% difference was measured in the Lower Valley. Since production at depth must have been identical for matching faces, these differences imply spatial variability in the surface production rates of cosmogenic nuclides after detachment. The variability in production rates is due to differences in topographic shielding and geometry of the boulders relative to their matching cliff faces. This variability should be accounted for by using proper correcting factors when calculating the exposure age of the sampled surfaces. Both in the Upper Valley and the Lower Valley ages of boulders and their matching cliff faces are similar within 1σ (Table 2, Fig. 1).

The common practice is to average the ages of the dated boulder and its matching cliff face to establish a time for the boulder detachment event. Thus, we would attribute an age of 3.1±1.1 ka for the "Closer Pile" in the Lower Valley (the average ¹⁰Be age of samples Timna 2 and Timna 3) and an age of 15.2±2.6 ka for the Upper Valley site (the average ¹⁰Be age of samples Timna 10 and Timna11). However, we have additional constraints on ages from these sites, and rather than average our calculated ¹⁰Be ages we can examine them in light of these constraints.

In the Lower Valley, OSL dating of sediments that accumulated behind the fallen boulders helped constrain the age of the boulder detachment event. Since the sediments accumulated behind the fallen boulders they must be younger than the boulders. The age relation between the dated boulder in the "Farther Pile" (15.1±2.1 ka) and the sediments in the pit upstream (Pit-1: 4.3±0.9 ka and 6.3±0.8 ka), support the field relations. Both sedimentary units that are exposed in the pit are younger than the boulder (Fig. 3). This is not the case in the "Closer Pile". The sedimentary units in the pit below the "Closer Pile" (Pit-2: 3.2±0.6 ka and 5.2±1.0 ka) are as old or older than the dated boulder (2.5±0.7 ka; Fig. 3), suggesting the boulder’s age is underestimated. Furthermore, the lower unit in pit 2 is even older than the average age of the boulder and its matching cliff face. The dated cliff (3.7±0.8 ka), although not older than lower unit in the pit, and although not entirely distinguishable from the boulder age, represents a more reasonable age for the boulder detachment since it is older than the upper sedimentary unit and, within 2 j, similar to the lower unit in pit 2.

It is apparent that the calculated scaling factor of 0.80, attributed to sample Timna 2, was not sufficient to enable a calculated exposure age that is as old or older than the OSL ages of the sedimentary units in pit 2. A scaling factor of 0.55 (or an actual surface production rate of 2.97 atoms g^-1 yr^-1) is required to calculate a boulder exposure age that is similar to the cliff face age. This production rate is 32% lower than the production rate we calculated for sample Timna 2 (4.33 atoms g^-1 yr^-1) based on the measured shielding parameters.

Combining the exposure ages determined from cosmogenic isotopes and the OSL ages in the three examined sites, the results suggest several boulder detachment events. The oldest one (31.0±4.9 ka) is recorded only in the third site by the age of the boulder that is located 20 meters from its source cliff. The second seems to be recorded in all three sites. The average age of samples Timna 10 and Timna 11 from the Upper Valley imply a detachment age of 15.1±3.4 ka, similar to the age of sample Timna 1 (15.1±2.1 ka) from the Lower Valley, and similar (within 1σ) to the age of the boulder situated 14 meters from the cliff in the third site and dated by OSL (13.9±4.8 ka). The third, and youngest event is recorded only at the "Closer Pile" in the Lower Valley (3.7±0.8 ka).

Rock falls and earthquakes

from Matmon et al. (2005:811-814)

Many processes, including tectonic, climatic, and environmental factors can cause rockslides [40]. Many of these factors can be eliminated in the case of rockslides in Timna. Snowmelt, freeze and thaw effects, ground water seepage, and tree root wedging can be ruled out due to the hot and hyperarid conditions in the area. Although rain storms and the result ing expansion of clay and salt particles in cracks is a plausible mechanism, we would expect boulders to detach one at a time rather than in groups that form large piles as is the case in Timna. Furthermore, the frequency of clay and salt wetting events is not high enough to generate proper stress in the fractures to allow boulder release. Several observations, mainly the large size of the boulders (many of the boulders have at least one dimension longer then 5 meters) and the agreement in ages of boulder piles in the sampling locations suggest that each pile of boulders was detached from its source cliff in a single event. The proximity of Timna to the DSF, where GT M6 earthquakes are common, suggests that ground shaking due to seismic events is the most likely cause for the Timna boulder slides.

A significant observation at the "Upper Valley" site supports the assumption that earthquakes are the major trigger of rock falls in the Timna area. A big block (5 m x 1.5 m x 2 m, point A in Fig. 4) is bounded by fractures. The fracture at the back of the boulder is about 30 cm wide and indicates the movement of the boulder in the direction of the cliff’s free face. The fracture at the base of the boulder is tilted towards the free face at an angle of less than 10° (Fig. 4). This angle is much lower than the maximum angle of repose of sandstone even when accounting for pore pressure effects [41]. Hence, the sliding of the block requires ground acceleration in a direction that is perpendicular to the cliff face to allow the opening of the back fracture. Ground acceleration in this desired direction is achieved by earthquakes generated along the adjacent segment of the DSF.

Fig. 4

Boulder pile at the "Upper Valley" location. Dated boulder is marked by dashed lines. Bedrock block (5 m x 1.5 m x 2 m; marked by the letter A) is bounded by fractures.
The back fracture of block A is open ~30 cm and suggests the movement of the boulder in the direction of the cliff’s free face.
Cross-section through the bedrock block perpendicular to the cliff face shows the gentle angle of the base fracture of the block. This angle (LT 10°) indicates that movement requires ground acceleration

Matmon et al. (2005)

Rock falls are sensitive recorders of strong ground motion resulting from earthquakes [42]. Synchronous rock falls may indicate the occurrence of past earthquakes and rock fall timing may constrain earthquake recurrence intervals and magnitude. The relation between seismic events and the formation of boulders in rock falls is well established. A worldwide correlation between landslide size and distribution and variables such as earthquake magnitude and the specific ground-motion characteristics was determined by [43]. A coseismic lichenometry model was developed in New Zealand following the discovery that lichens growing on rocky hill slopes recorded synchronous pulses of rock falls generated by historical earthuakes. The lichenometry model was used to date boulders and rock falls associated with earthquakes [44–48].

Current measurements in the southern Arava Valley along the DSF system show no seismic activity [49]. However, historical evidence documents several large seismic events [50,51]. Paleoseismic studies in the southern Arava Valley suggest that late Pleistocene earthquakes ranged in magnitude between 6.7 and 7.1 and the average recurrence interval was 2.8±0.7 ky [52]. These studies indicate that Holocene earthquakes were more frequent, with an average recurrence interval of 1.2±0.3 ky, but with smaller magnitudes that ranged between M5.9 and M6.7. Several studies in the northern Arava Valley also suggest frequent Holocene and late Pleistocene seismic activity [4,50,53–56].

The time interval between the three boulder forming events recorded in this study is much longer than the recurrence interval of GT M6 earth quakes in the region. However, this discrepancy does not rule out seismic motion as the mechanism for the boulder formation. A possible explanation is that the interval between boulder forming events represents the time that is necessary for fractures to develop to the point of minimum friction between boulders and bedrock. During this time, earthquakes occur and gradually enhance the open ing of fractures that surround boulders. Once these fractures are sufficiently developed, the next major earthquake releases the boulders. The spacing between the fractures determines the thickness of collapsed wall during each rock fall event and the level of sandstone lithification determines the resistance of the fresh exposed rock to weathering. Continued study of additional rock piles at Timna may help to better constrain the temporal frequency of rock fall events, and improve the correlation between rock falls and earthquakes.

Rates of cliff retreat

from Matmon et al. (2005:814-815)

The unweathered sandstone boulder and cliff faces at Timna suggest that retreat occurs mostly during discrete rock fall events that are separated by thousands of years while continuous grain-by-grain weathering is minor. Meaningful cliff retreat rates can be obtained if enough collapse cycles are recorded. Nevertheless, cliff retreat rates estimated from only several recorded cycles can provide a general and basic knowledge as to the rate of landscape development in the area. The record we obtained includes at the most three rock fall cycles, at 31 ka, 15 ka, and ~4 ka.

In the Upper Valley and at the "Closer Pile" in the Lower Valley, the thickness of the dated boulders represent the amount of retreat. The average thickness of the dated boulder (Timna 10) in the Upper Valley is 2.2 meters, implying a retreat rate of 0.14±0.02 m ky¹. At the "Closer Pile" in the Lower Valley the average thickness of the dated boulder (Timna 2) is 8 meters implying a retreat rate of 2±0.2 m ky¹. However, in both cases, retreat rates are not based on a complete rock fall cycle (from one rock fall event to the next rock fall event) and therefore, are maximum rates at each location. Better estimates of the average cliff retreat rates can be calculated from complete cycles.

In the Lower Valley, the "Farther Pile" was deposited 15.1±2.1 ka (sample Timna 1). There is no direct evidence for the thickness of the collapsed section that formed the "Farther Pile". However, we assume that the spacing of cracks that determine the thickness of the collapsed sections is similar within each site, and therefore, the collapsed section that formed the "Farther Pile" was about 8 meters thick, similar to the "Closer Pile" thickness. The time interval between the "Farther Pile" and the "Closer Pile" is 11.4±2.2 ky. This time interval implies a cliff retreat rate of 0.7±0.1 m ky¹.

In the third site, we can distinguish two complete cycles of repeated rock falls. One is between the time of collapse of boulder TMN-7 and the time of collapse of boulder TMN-8. The distance between boulders TMN-7 and TMN-8 is 6 meters and the age difference is 17.1±6.9 ky implying a cliff retreat rate of 0.4±0.16 m ky^-1. The other complete cycle is between the time of collapse of boulder TMN-8 and the present. We assume that boulder TMN-9 (which was not dated) has fallen only recently and that the distance it has fallen from the cliff (4 meters) represents the typical distance that boulders fall from the source cliff. Thus, the age of boulder TMN-8 (13.9±4.8 ka) and the distance from its source cliff (the recent 14 meters minus 4 meters of original distance) imply a cliff retreat rate of 0.7±0.2 m ky^-1. The weathered state of the boulders in the third site compared to the fresh boulders in the Upper and Lower Valleys and the existence of only single boulders in the third site as opposed to boulder piles in the other site can be explained by rocks less resistant to chemical weathering.

Cliff retreat rates determined from the rock falls in Timna range between 0.14 and 2 m ky^-1. A more constrained range of 0.4 to 0.7 m ky^-1 is calculated from the complete collapse cycles. Similar cliff retreat rates that range between 0.1 and 0.85 m ky^-1 were calculated in arid environments, [57–60].

In Timna, rates of cliff retreat are faster by an order of magnitude than the rate that the upper surfaces are lowered by weathering (determined from samples Timna 20 and Timna 21 - 24.3±3.6 mm ky^-1). Even though cliffs retreat through the episodic process of rock fall events while the upper surface weather continuously, over time the difference in rates maintains the sharp morphology of Timna and vertical slopes (cliffs) do not roll back to become more gentle and rounded slopes. Cliff retreat obviously initiated when voids large enough opened and enabled the collapse process to begin. This space opening was probably first accomplished through incision of stream channels through the sandstone following the development of the Arava Valley topographical base level. Assuming the range of cliff retreat rates calculated in this study, the 30 m wide Lower Valley opened approximately 40- 70 ka. This range of ages determines the time at which the upstream-moving signal of the lowering base level of the DSF reached the Lower Valley site.

Conclusions

from Matmon et al. (2005:815-816)

The combined ages from the three investigated sites suggest that the boulders were deposited in three events at 31 ka, 15 ka, and 4 ka. The most likely mechanism for boulder formation in this tectonically active hyperarid region is by tectonically induced ground motion. Paleoseismic studies in the area suggest a recurrence interval of 1000–2000 yr for earthquakes GT M6. The cosmogenic and OSL age dating of the boulders suggests that boulder-formation events do not occur as frequently as GT 6 M earthquakes occur in this area Therefore, it is apparent that boulder falls in the study area do not represent a full seismic record. We suggest that boulder formation events occur during large earthquakes only after blocks of rock are sufficiently separated from bedrock by large and well developed cracks. These cracks allow the detachment of the boulders when the next earthquake occurs.

Cliff retreat rates determined from the rock falls in Timna range between 0.14 and 2 m ky¹. A more constrained range of 0.4 to 0.7 m ky¹ is calculated from the complete collapse cycles. These retreat rates are similar to those calculated in other arid regions. Field observations suggest that the retreat occurs mainly during rock fall events and that continuous weathering of the rock faces is less important. Since cliff retreat rates in Timna are an order of magnitude larger than the calculated rate of surface lowering, as calculated in the Lower Valley site, the sharp and imposing landscape characteristic to this region is preserved.

We described the spatial and temporal relations between boulders and their original in-situ locations on sandstone bedrock cliffs. In each sampled pair, a boulder and its matching cliff face yielded significantly different ¹⁰Be concentration. A 10–20% difference in ¹⁰Be concentration between boulders and their matching cliff faces imply a spatial variability in cosmogenic nuclide production rates. Differences in cosmogenic nuclide production rates arise from differences in topographic shielding and boulder geometry. The variability in shielding factors affects each boulder and its matching cliff face differently after detachment. Calculated mean ages differed by 14-32%. However, the uncertainties associated with the age calculations prevent us from distinguishing between boulder and matching cliff ages. Age calculations considered the pre-detachment accumulation of ¹⁰Be produced by deep penetrating muons. The agreement between the ¹⁰Be and ¹⁴C model ages in sample Timna 2 implies that the estimated pre-detachment ¹⁰Be concentration was correct.

In the Lower Valley, field relations between the dated boulders and sediments that accumulated upstream behind them, dated using OSL, helped to constrain the formation age of the "Closer Pile", and suggested that the older exposure age of the cliff (3.7±0.8 ka) is a more reasonable age for the boulder detachment event. In case of complicated morphological setting it is useful to combine external constraints such as field relations or other dating methods to evaluate the reliability of cosmogenic exposure ages.

Shehoret Rockfalls - Seismic Unzipping (?) between 2000 and 3000 BCE (4-5 ka)

Discussion

Rinat et al. (2014) dated rockfalls between 3.6 ± 0.8 ka and 4.7 ± 0.7 ka from four sites (SH1, SH2, SH4, and SH10) in the vicinity of Nahal Shehoret (נחל שחודת). They suggested that if one considers the hyper-arid environment surrounding Nahal Shehoret, it is most likely that the dated rockfalls were triggered by earthquakes. After comparison with other dated rockfalls in the region and paleoseismic studies along the the entire length of the DSF system, they suggested that all these records combined imply intensive seismic activity around 4–5 ka along the DSF system and its main branch (the Carmel fault). Although they surmised that such intensive seismic activity could be explained by a single, large-scale (≥8 M) event which affected the entire length (~500 km) of the DSF and its margins, repeated smaller seismic events occurring throughout the entire length of the DSF and its main branch (the Carmel fault) was more likely because an 8 M earthquake most likely would have left its impression on the landscape (Keefer, 2002; Dai et al., 2011) and no evidence was found along the DSF system for such a large-scale event.

References

Rinat et al. (2014)

Abstract

from Rinat et al. (2014:260)

Rockfall ages in tectonically active regions provide information regarding frequency and magnitude of earth quakes. In the hyper-arid environment of the Dead Sea fault (DSF), southern Israel, rockfalls are most probably triggered by earthquakes. We dated rockfalls along the western margin of the DSF using terrestrial cosmogenic nuclides (TCN). At each rockfall site, samples were collected from simultaneously exposed conjugate boulders and cliff surfaces. Such conjugate samples initially had identical pre-fall (“inherited”) TCN concentrations. After boulder detachment, these surfaces were dosed by different production rates due to differences in post-fall shielding and geometry. However, in our study area, pre-rockfall inheritance and post-rockfall production rates of TCN cannot be evaluated. Therefore, we developed a numerical approach and demonstrated a way to overcome the above-mentioned problems. This approach can be applied in other settings where rockfalls cannot be dated by simple exposure dating. Results suggest rockfall ages between 3.6 ± 0.8 and 4.7 ± 0.7 ka. OSL ages of sediment accumulated behind the boulders range between 0.6 ± 0.1 and 3.4 ± 1.4 ka and support the TCN results. Our ages agree with dated earthquakes determined in paleoseismic studies along the entire length of the DSF and support the observation of intensive earthquake activity around 4–5ka.

Introduction

from Rinat et al. (2014:260-261)

Rockfalls are a typical mass wasting process that occurs when an entire rock wall, or individual blocks fall during a single event (Ahnert, 1996; Hantz et al., 2003). Tectonic, climatic, environmental, and random factors contribute to the occurrence of rockfalls. Nevertheless, the triggering cause for a rockfall is often a minor event such as a small earthquake, lightning, heavy rainfall, or human intervention (Schumm and Chorley, 1964; Ahnert, 1996; Matsuoka and Sakai, 1999; Marzorati et al., 2002).

Rockfalls are common features in tectonically active regions and serve as sensitive recorders of strong ground motion resulting from earthquakes (Bull and Menges, 1977; Zellmer, 1987; Bull et al., 1994; Vidrih et al., 2001). Synchronous rockfalls may indicate the occurrence of past earthquakes and rockfall timing may constrain earthquake recurrence intervals. The relation between seismic events and the detachment of boulders in rockfalls is well established (e.g., Keefer, 2002; Dai et al.,2011; Xu et al., 2013). A worldwide correlation between landslides and variables such as earthquake magnitude and groundmotion characteristics was determined by Keefer (1984). Moreover, in some regions, such as the southwest USA and New Zealand, a correlation between independently dated earthquakes and rockfalls has been demonstrated (e.g., Bull, 1996a,b). Rockfalls are also an important process controlling scarp retreat rates (e.g., Matmon et al., 2005). Thus, they are a dominant process in the development of landscapes in tectonically active regions.

The Dead Sea fault (DSF) is part of the African–Arabian plate boundary (Garfunkel, 1981; Garfunkel et al., 1981). It has been active since the Miocene (Bartov et al., 1980) and seismic activity is recorded along it continuously (Marco et al., 1996; Ken-Tor et al., 2001; Amit et al., 2002; Hofstetter, 2003; Kagan et al., 2005; Zilberman et al., 2005; Baer et al., 2008). Rockfalls are abundant along both the eastern and western margins of the DSF and are most likely associated with the repeated earthquakes that occur in the region. The main advantage in considering rockfalls as paleoseismic indicators is the fact that they are not associated with a specific fault. Thus, the temporal distribution of rockfall events is a recorder to the overall seismic activity in a region.

In this study we date rockfalls located along the western margin of the southern DSF using cosmogenic isotope measurements. Dating such rockfalls contributes to our knowledge of geomorphological processes and to the understanding of the overall long-term seismic activity in the southern DSF area. The calculation of simple exposure ages of rockfalls is not applicable in our study site. Numerical procedure is required to calculate rockfall ages, which range between 3.6 ± 0.8 and 4.7 ± 0.7 ka in our study sites. These ages agree well with other boulder piles dated in the region (Matmon et al., 2005) and with earthquake ages determined in paleoseismic studies along the entire length of the DSF system (Marco et al., 1996; Gluck, 2001; Ken-Tor et al., 2001; Amit et al., 2002; Hofstetter, 2003; Migowski et al., 2004; Kagan et al., 2005; Marco et al., 2006; Baer et al., 2008; Zilberman et al., 2008; Braun et al., 2010; Katz et al., 2010; Shaked et al., 2011). All compiled paleoseismic data suggest intensive earthquake activity along the entire DSF system and its main branch (the Carmel fault) around 4–5 ka.

Research Site

from Rinat et al. (2014:261)

The research area is located on the western margin of the southern Arava Valley, DSF, Israel (Fig. 1). This extremely arid environment is subject to rare rain events and very low precipitation (~30 mm/yr; Israel Meteorological Service). Sub-zero temperatures are rarely reached (Goldreich and Karni, 2001; Israel Meteorological Service). Cambrian sandstones are exposed in the study area and its vicinity, composed of ~90% quartz grains (Karcz et al., 1971; Weissbrod, 1980; Druckman et al., 1993). The sandstones overlie Precambrian crystalline basement and form steep cliffs (Avigad et al., 2005). Field observations indicate that two processes dominate the research site morphology: cliff retreat and lowering of the upper surface above each rockfall site. Cliff retreat is dominated by rockfalls, while the lowering of the upper surface is dominated by slow erosion of the leveled surface above each site. Piles of huge (~120 m³) sandstone boulders (Fig. 2), which are the subject of this research, are found in many locations at the base of these cliffs. Boulder faces and detachment scars on the cliff are virtually intact. In specific sites the boulders can be traced, within ~1-cm accuracy, back to their original location on the cliff (Fig. 3). The minor erosion enables the identification of multiple boulder and cliff faces that were exposed at the same time, and thus several samples with identical exposure time can be collected. Post-fall geometry of each sampled face (e.g., shielding and surface tilt) controls the differences in production rates at the various sampling locations, which otherwise share similar altitudes, latitudes, and rock composition. Furthermore, selected boulders are positioned on flat surfaces, so the possibility of post-detachment movement, and time-dependent geometrical scaling factors, is decreased.

We surveyed a large area in which Cambrian sandstone cliffs dominate the landscape and mapped many rockfalls. Several features were examined in each rockfall. Boulders had to be sufficiently large (several meters in each dimension). Their orientation and location, prior to the fall, had to be easily determined and as many pair samples as possible (boulder–boulder, or boulder–cliff) were needed. The sample face, before detachment, had to be shielded by several meters of rock to reduce isotope accumulation by high-energy neutrons and muons prior to detachment. The boulders' current position had to be on a flat surface, so the possibility of post-detachment movement can be neglected. Ultimately, we focused our study on four rockfalls, SH1, SH2, SH4, and SH10, in the vicinity of Nahal Shehoret (Fig. 2), which best met the above-mentioned criteria.

Site SH1 includes two boulders located at the foot of their original cliff. In site SH2 there are three boulders that once formed the cliff's wall. Site SH4 includes two boulders located at the foot of their original cliff. At site SH10, a pile of five boulders is located at the foot of their original cliff. All boulders were matched to their corresponding cliff faces.

Implications for earthquakes along the DSF

from Rinat et al. (2014:271-272)

In the hyper-arid environment of the southwestern Arava the most likely trigger for rockfalls is the occurrence of seismic events (earthquakes). The area is nearly devoid of vegetation. Rain events are rare and sub-zero temperatures rarely occur. Thus, tree root wedging and freeze–thaw cycles are ruled out as probable rockfall triggers. On the other hand, seismic activity is continuously recorded and earthquakes occur frequently.

The obtained rockfall ages from our research sites range between 3.6 ± 0.8 and 4.7 ± 0.7 ka, and are identical within their uncertainty limits. A mutual age for all four sites is calculated to 4.3 ± 1 ka (Fig. 9). These rockfalls were most likely formed during closely spaced earthquakes or possibly during a single seismic event; the resolution of the dating method does not enable us to distinguish between these alternatives. Apparently, many paleoseismic records along the DSF and its margins indicate the occurrence of earthquakes during this time period (Fig. 10). These studies employ various dating methods and investigate a large range of paleoseismic archives. For example, Matmon et al. (2005) dated three boulder piles in the Timna Valley (~17 km north of our study site). Rockfall ages in Timna were recalculated using the presently accepted production rate and Dunai (2000, 2001) scaling scheme to allow comparison with our results. One Holocene rockfall yielded an average age of 4.5 ± 1.0 ka. Amit et al. (2002) analyzed fault lines in the vicinity of the Shehoret alluvial fan (~5 km from the research site) and identified Holocene seismic events with an estimated magnitude between M 5.9 and M 6.7 and a recurrence interval of 1.2 ± 0.3 ka. In that study, Holocene seismic events cluster in several groups of which one is at ~5 ka. Shaked et al. (2011) dated a buried coral reef at the northern edge of the Gulf of Aqaba (~15 km south of our study site) to ~4.7 ka, using radiocarbon and U-series. The reef was found in a pristine condition. Its subsidence and burial are attributed to down faulting caused by a seismic event that occurred at that time and caused the burial of that part of the reef.

Earthquake records from more distant locations along the Dead Sea rift and its main branching fault (the Carmel Fault) also indicate the occurrence of earthquakes at 4–5 ka (e.g., Gluck, 2001; Amit et al., 2002; Migowski et al., 2004; Kagan et al., 2005; Matmon et al., 2005; Marco et al., 2006; Zilberman et al., 2008; Braun et al., 2010; Katz et al., 2010; Shaked et al., 2011; Fig. 10). Although, some of the above-mentioned records suggest repeated seismic events throughout the Holocene, the only time period in which all records indicate seismic events is the interval between 4 and 5 ka. The widespread occurrence of seismic events at 5 ka has been considered and discussed by Braun et al. (2010) but only for the northern part of the DSF. Here we suggest that the events spanned the entire length of the DSF.

The evidence of simultaneous events at ~5 ka along the DSF system could be explained in two ways. Possibly, a single, large-scale (≥8 M) event affected the entire length (~500 km) of the DSF and its margins. Alternatively, it is possible that repeated, smaller seismic events occurred throughout the entire length of the DSF and its main branch (the Carmel fault). Within the resolution of the dating methods, it is not possible to distinguish one alternative from the other. However, an 8 M earthquake most likely would have left its impression on the landscape (Keefer, 2002; Dai et al., 2011) and no evidence was found along the DSF system for such a large-scale event. Furthermore, paleoseismic studies conducted along the DSF system generally estimate earthquake magnitudes at 6–7 M. Therefore, it is more likely that the seismic events seen in the various records along the length of the DSF are the result of closely timed earthquakes that occurred between 4 and 5 ka. Earthquake clustering on millennia time scale has also been inferred for the DSF during the late Pleistocene (Marco et al., 1996) and the Holocene (Migowski et al., 2004). Yet these studies were confined to seismites in the close proximity of the Dead Sea. Such fault behavior is not unique to the DSF and has been observed on other major strike slip faults such as the San Andreas, California, the Denali fault, Alaska, and the North Anatolian fault, Turkey (e.g., Prentice, 1999; Lettis et al., 2002; Frankel, 2004).

Conclusions

from Rinat et al. (2014:272)

In this study we dated, using measured concentrations of cosmogenic ¹⁰Be, four rockfalls located along the western margin of the southern DSF. The calculation of simple exposure ages of rockfalls was not applicable as a result of the complicated geometric setting of the dated boulder and cliff faces and the apparent complex exposure–burial history of the boulders. Therefore, both pre-rockfall inheritance and post-rockfall production rates could not be evaluated. A numerical procedure was demonstrated to overcome the above-mentioned problems and date boulder piles. The results suggest numerical estimated boulder-fall ages that range between 3.6 ± 0.8 ka and 4.7 ± 0.7 ka. Optically stimulated luminescence ages of sediment accumulated behind the fallen boulders range between 0.6 ± 0.1 ka and 3.4 ± 1.4 ka and support the TCN results. It is recommended to use this model with no less than four samples, all with identical exposure time, and, preferably, at different pre-fall depths as the uncertainty associated with the derived exposure age decreases as the number of samples increases.

Considering the hyper-arid environment it is most likely that the dated rockfalls were triggered by earthquakes. Therefore, our rockfalls ages contribute to the paleoseismic record of the DSF. Our ages agree well with other rockfalls dated in the region and with earthquake ages determined in paleoseismic studies along the entire length of the DSF system. We suggest that the entire suit of records implies intensive seismic activity around 4–5 ka along the DSF system and its main branch (the Carmel fault).

Closer Pile Lower Timna Valley Earthquake - 2500-900 BCE (3.7±0.8 ka, 1700 BCE ± 800)

Discussion

Matmon et al. (2005:811) used the combined exposure ages determined from cosmogenic isotopes and OSL ages from three rockfall sites at Timna to suggest several boulder detachment events which they surmise were caused by earthquakes. They only dated one event in the Holocene. This event was dated to 3.7±0.8 ka based on the "Closer Pile" in the Lower Valley. Agnon (2014:232) discussed the same event but listed a slightly different date - 32-4.5 ka.

References

Matmon et al. (2005)

Abstract

from Matmon et al. (2005:803-804)

In this study, we explored the spatial and temporal relations between boulders and their original in-situ locations on sandstone bedrock cliffs. This was accomplished by combining field observations with dating methods using cosmogenic isotopes (¹⁰Be and ¹⁴C ) and optically stimulated luminescence (OSL). Our conclusions bear both on the landscape evolution and cliff retreat process in the hyperarid region of Timna and on the methodology of estimating exposure ages using cosmogenic isotopes.

We recognize three discrete rock fall events, at 31 ka, 15 ka, and 4 ka. In this hyperarid region, the most plausible triggering mechanism for rock fall events is strong ground acceleration caused by earthquakes generated by the nearby Dead Sea fault (DSF). Our record, however, under represents the regional earthquake record implying that ongoing development of detachment cracks prior to the triggering event might be slower than the earthquake cycle.

Cliff retreat rates calculated using the timing of rock fall events and estimated thickness of rock removed in each event range between 0.14 m ky⁻¹ and 2 m ky⁻¹. When only full cycles are considered, we derive a more realistic range of 0.4 m ky⁻¹ to 0.7 m ky⁻¹. These rates are an order of magnitude faster than the calculated rate of surface lowering in the area. We conclude that sandstone cliffs at Timna retreat through episodic rock fall events that preserve the sharp, imposing, landscape characteristic to this region and that ongoing weathering of the cliff faces is minor.

A 10%–20% difference in the ¹⁰Be concentrations in samples from matching boulder and cliff faces that have identical exposure histories and are located only a few meters apart indicates that cosmogenic nuclide production rates are sensitive to shielding and vary spatially over short distances. However, uncertainties associated with age calculations yielded boulder and matching cliff face ages that are similar within 1σ . The use of external constraints in the form of field relations and OSL dating helped to establish each pair's age. The agreement between calculated ¹⁴C and ¹⁰Be ages indicates that the accumulation of ¹⁰Be at depth by the capture of slow deep-penetrating muons was properly accounted for in the study.

Introduction

from Matmon et al. (2005:804-805)

Cliffs and unstable slopes are sources of detritus for a variety of mass-movement processes. Rock falls occur when recently detached segments of bedrock slide down slope. This is likely to occur along steep mountain sides and may have catastrophic results. A detached block falls from a rock wall when the gravity-determined shear stress is greater than the shearing resistance [1]. The detachment fissure separating the block from the outcrop is usually already present.

Tectonic, climatic, environmental, and random factors contribute to the occurrence of rockslides. Many rock falls are caused by undercutting and steepening of the rock wall because of lateral stream erosion or accelerated back weathering of the wall foot. Nevertheless, the triggering of a block fall or a rock fall after an extended period of preparation is often a minor event such as a small earthquake, lightning, or heavy rainfall [1]. The rate at which rock falls occur may determine the rate of cliff retreat and the potential hazard associated with these falls. Thus, dating past rock falls can help to better understand the process of cliff retreat. To estimate the rate of cliff retreat, it is essential to determine the position of the cliff at some time in the past and examine the processes by which it arrived at its present position.

Dating boulders, mainly with cosmogenic radio nuclides (CRN), has become common practice in geomorphologic studies. On active moraines and alluvial fans, deposited boulders might contain CRN from prior exposure and might be buried, re-exposed and rolled. Thus, boulder ages represent the time at which they become stable elements in the landscape [2,3]. The time between deposition and stabilization might be in the order of 10²–10³ years. In the case of rock falls, boulders are deposited instantaneously. Previously shielded rock faces are exposed, and their ages correlate to discrete geomorphic events. In arid environments, where erosion rates are slow, fresh and well-preserved exposed rock faces represent the actual time of the rock fall events. Rock piles may indicate the general past position of a cliff. Thus, we hypothesize that by dating rock piles, or individual boulders in such piles, we open a window that enables us to better understand landscape development in arid environments controlled by rock falls and other cliff retreat processes.

In the present study, we describe the spatial and temporal relations between very large angular blocks (which will be referred to as boulders) and their original in-situ locations on sandstone bedrock cliffs along the western margins of the DSF, southern Israel (Fig. 1). Boulder detachment events were dated by combining exposure age dating using cosmogenic isotopes (¹⁰Be, ¹⁴C) and by OSL. The results indicate that boulder detachment events in this area occur less often than GT 6 M earthquakes [4]. Rock falls occur only when boulders are sufficiently detached from the bedrock cliffs by large and developed cracks and are thus susceptible to the next triggering event.

Geological and morphological setting

from Matmon et al. (2005:805-806)

Timna is located at the western margin of the southern Arava Valley, a hyperarid region extending from the Dead Sea basin to the Gulf of Aqaba (Fig. 1). Timna is an uplifted core of late Precambrian igneous rocks [5] overlain by Cambrian and Lower Cretaceous clastic sediments. It is a bowl-shaped valley, about 8 km in diameter, and is surrounded by cliffs of Upper Cretaceous limestone and dolomite rising up to 600 m above the valley floor.

Landscape development in the Timna area is controlled by the tectonic activity along the southern section of the DSF (Hannan Ginat, personal communication). Erosional truncation of the Upper Cretaceous carbonate cap rocks during the Oligocene and Miocene exposed the more erodable Lower Cretaceous sandstone. As the southern Arava Valley developed into a topographic base level, the sandstone was eroded and transported towards that base level, the Timna Valley was incised, and the Precambrian basement and Cambrian sediments were exposed at the base of the Timna Valley.

The Cambrian sandstone in Timna reaches a thickness of about 100 m [5] and forms cliffs that dominate the landscape. Within the Cambrian sandstone terrain, erosion is dominated both by detachment of massive boulders from the sandstone cliffs and by slow weathering of the cliff faces, as evident from 0.1-m-scale cavities (tafoni). Generally, boulders detached from sandstone cliffs are found in piles and are usually coated with varnish. In many locations, the cliff face and the faces on the boulders match perfectly both in detail and overall morphology. Some of these boulders can be accurately traced to their original position on the cliff from which they were detached (Fig. 2).

We concentrated on three boulder sites in the Timna area (Fig. 1). The first, the Lower Valley, contains two boulder piles (one close to the source sandstone cliff face, which we call the "Closer Pile", and one farther away, the "Farther Pile") detached from a cliff that rises 10–14 m above the valley floor. The boulders at this site were deposited at the mouth of the Lower Valley (30 x 60 m), creating a natural dam, and causing upstream accumulation of sandy sediments on the valley floor. The second site, the Upper Valley, contains several boulder piles. We concentrated on a single boulder pile detached from a nearby 10 m cliff. At both locations boulders are not significantly affected by erosion and perfectly match the cliff faces (Fig. 2). Unlike the first two sites that contain unweathered boulder piles, the third site we examined contains three weathered boulders that are located at distances of 20, 14, and 4 m from the source cliff (Fig. 1E).

Discussion

Boulder ages and boulder detachment events

from Matmon et al. (2005:810-811)

Both in the Upper and Lower Valley sites, a boulder and its matching cliff face were sampled. In both cases, the sampled boulder can be matched perfectly back to its original location on the cliff. The assumed history of the boulders and their matching cliff faces imply an identical exposure age for each pair. After detachment, the boulders differed from the cliff faces in their topographic shielding and geometric properties. Therefore, production rates of cosmogenic nuclides in the boulders were different from Valley, a 10% difference in ¹⁰Be concentration between the boulder and its matching cliff face was measured; a 20% difference was measured in the Lower Valley. Since production at depth must have been identical for matching faces, these differences imply spatial variability in the surface production rates of cosmogenic nuclides after detachment. The variability in production rates is due to differences in topographic shielding and geometry of the boulders relative to their matching cliff faces. This variability should be accounted for by using proper correcting factors when calculating the exposure age of the sampled surfaces. Both in the Upper Valley and the Lower Valley ages of boulders and their matching cliff faces are similar within 1σ (Table 2, Fig. 1).

The common practice is to average the ages of the dated boulder and its matching cliff face to establish a time for the boulder detachment event. Thus, we would attribute an age of 3.1±1.1 ka for the "Closer Pile" in the Lower Valley (the average ¹⁰Be age of samples Timna 2 and Timna 3) and an age of 15.2±2.6 ka for the Upper Valley site (the average ¹⁰Be age of samples Timna 10 and Timna11). However, we have additional constraints on ages from these sites, and rather than average our calculated ¹⁰Be ages we can examine them in light of these constraints.

In the Lower Valley, OSL dating of sediments that accumulated behind the fallen boulders helped constrain the age of the boulder detachment event. Since the sediments accumulated behind the fallen boulders they must be younger than the boulders. The age relation between the dated boulder in the "Farther Pile" (15.1±2.1 ka) and the sediments in the pit upstream (Pit-1: 4.3±0.9 ka and 6.3±0.8 ka), support the field relations. Both sedimentary units that are exposed in the pit are younger than the boulder (Fig. 3). This is not the case in the "Closer Pile". The sedimentary units in the pit below the "Closer Pile" (Pit-2: 3.2±0.6 ka and 5.2±1.0 ka) are as old or older than the dated boulder (2.5±0.7 ka; Fig. 3), suggesting the boulder’s age is underestimated. Furthermore, the lower unit in pit 2 is even older than the average age of the boulder and its matching cliff face. The dated cliff (3.7±0.8 ka), although not older than lower unit in the pit, and although not entirely distinguishable from the boulder age, represents a more reasonable age for the boulder detachment since it is older than the upper sedimentary unit and, within 2 j, similar to the lower unit in pit 2.

It is apparent that the calculated scaling factor of 0.80, attributed to sample Timna 2, was not sufficient to enable a calculated exposure age that is as old or older than the OSL ages of the sedimentary units in pit 2. A scaling factor of 0.55 (or an actual surface production rate of 2.97 atoms g^-1 yr^-1) is required to calculate a boulder exposure age that is similar to the cliff face age. This production rate is 32% lower than the production rate we calculated for sample Timna 2 (4.33 atoms g^-1 yr^-1) based on the measured shielding parameters.

Combining the exposure ages determined from cosmogenic isotopes and the OSL ages in the three examined sites, the results suggest several boulder detachment events. The oldest one (31.0±4.9 ka) is recorded only in the third site by the age of the boulder that is located 20 meters from its source cliff. The second seems to be recorded in all three sites. The average age of samples Timna 10 and Timna 11 from the Upper Valley imply a detachment age of 15.1±3.4 ka, similar to the age of sample Timna 1 (15.1±2.1 ka) from the Lower Valley, and similar (within 1σ) to the age of the boulder situated 14 meters from the cliff in the third site and dated by OSL (13.9±4.8 ka). The third, and youngest event is recorded only at the "Closer Pile" in the Lower Valley (3.7±0.8 ka).

Rock falls and earthquakes

from Matmon et al. (2005:811-814)

Many processes, including tectonic, climatic, and environmental factors can cause rockslides [40]. Many of these factors can be eliminated in the case of rockslides in Timna. Snowmelt, freeze and thaw effects, ground water seepage, and tree root wedging can be ruled out due to the hot and hyperarid conditions in the area. Although rain storms and the result ing expansion of clay and salt particles in cracks is a plausible mechanism, we would expect boulders to detach one at a time rather than in groups that form large piles as is the case in Timna. Furthermore, the frequency of clay and salt wetting events is not high enough to generate proper stress in the fractures to allow boulder release. Several observations, mainly the large size of the boulders (many of the boulders have at least one dimension longer then 5 meters) and the agreement in ages of boulder piles in the sampling locations suggest that each pile of boulders was detached from its source cliff in a single event. The proximity of Timna to the DSF, where GT M6 earthquakes are common, suggests that ground shaking due to seismic events is the most likely cause for the Timna boulder slides.

A significant observation at the "Upper Valley" site supports the assumption that earthquakes are the major trigger of rock falls in the Timna area. A big block (5 m x 1.5 m x 2 m, point A in Fig. 4) is bounded by fractures. The fracture at the back of the boulder is about 30 cm wide and indicates the movement of the boulder in the direction of the cliff’s free face. The fracture at the base of the boulder is tilted towards the free face at an angle of less than 10° (Fig. 4). This angle is much lower than the maximum angle of repose of sandstone even when accounting for pore pressure effects [41]. Hence, the sliding of the block requires ground acceleration in a direction that is perpendicular to the cliff face to allow the opening of the back fracture. Ground acceleration in this desired direction is achieved by earthquakes generated along the adjacent segment of the DSF.

Fig. 4

Boulder pile at the "Upper Valley" location. Dated boulder is marked by dashed lines. Bedrock block (5 m x 1.5 m x 2 m; marked by the letter A) is bounded by fractures.
The back fracture of block A is open ~30 cm and suggests the movement of the boulder in the direction of the cliff’s free face.
Cross-section through the bedrock block perpendicular to the cliff face shows the gentle angle of the base fracture of the block. This angle (LT 10°) indicates that movement requires ground acceleration

Matmon et al. (2005)

Rock falls are sensitive recorders of strong ground motion resulting from earthquakes [42]. Synchronous rock falls may indicate the occurrence of past earthquakes and rock fall timing may constrain earthquake recurrence intervals and magnitude. The relation between seismic events and the formation of boulders in rock falls is well established. A worldwide correlation between landslide size and distribution and variables such as earthquake magnitude and the specific ground-motion characteristics was determined by [43]. A coseismic lichenometry model was developed in New Zealand following the discovery that lichens growing on rocky hill slopes recorded synchronous pulses of rock falls generated by historical earthuakes. The lichenometry model was used to date boulders and rock falls associated with earthquakes [44–48].

Current measurements in the southern Arava Valley along the DSF system show no seismic activity [49]. However, historical evidence documents several large seismic events [50,51]. Paleoseismic studies in the southern Arava Valley suggest that late Pleistocene earthquakes ranged in magnitude between 6.7 and 7.1 and the average recurrence interval was 2.8±0.7 ky [52]. These studies indicate that Holocene earthquakes were more frequent, with an average recurrence interval of 1.2±0.3 ky, but with smaller magnitudes that ranged between M5.9 and M6.7. Several studies in the northern Arava Valley also suggest frequent Holocene and late Pleistocene seismic activity [4,50,53–56].

The time interval between the three boulder forming events recorded in this study is much longer than the recurrence interval of GT M6 earth quakes in the region. However, this discrepancy does not rule out seismic motion as the mechanism for the boulder formation. A possible explanation is that the interval between boulder forming events represents the time that is necessary for fractures to develop to the point of minimum friction between boulders and bedrock. During this time, earthquakes occur and gradually enhance the open ing of fractures that surround boulders. Once these fractures are sufficiently developed, the next major earthquake releases the boulders. The spacing between the fractures determines the thickness of collapsed wall during each rock fall event and the level of sandstone lithification determines the resistance of the fresh exposed rock to weathering. Continued study of additional rock piles at Timna may help to better constrain the temporal frequency of rock fall events, and improve the correlation between rock falls and earthquakes.

Rates of cliff retreat

from Matmon et al. (2005:814-815)

The unweathered sandstone boulder and cliff faces at Timna suggest that retreat occurs mostly during discrete rock fall events that are separated by thousands of years while continuous grain-by-grain weathering is minor. Meaningful cliff retreat rates can be obtained if enough collapse cycles are recorded. Nevertheless, cliff retreat rates estimated from only several recorded cycles can provide a general and basic knowledge as to the rate of landscape development in the area. The record we obtained includes at the most three rock fall cycles, at 31 ka, 15 ka, and ~4 ka.

In the Upper Valley and at the "Closer Pile" in the Lower Valley, the thickness of the dated boulders represent the amount of retreat. The average thickness of the dated boulder (Timna 10) in the Upper Valley is 2.2 meters, implying a retreat rate of 0.14±0.02 m ky¹. At the "Closer Pile" in the Lower Valley the average thickness of the dated boulder (Timna 2) is 8 meters implying a retreat rate of 2±0.2 m ky¹. However, in both cases, retreat rates are not based on a complete rock fall cycle (from one rock fall event to the next rock fall event) and therefore, are maximum rates at each location. Better estimates of the average cliff retreat rates can be calculated from complete cycles.

In the Lower Valley, the "Farther Pile" was deposited 15.1±2.1 ka (sample Timna 1). There is no direct evidence for the thickness of the collapsed section that formed the "Farther Pile". However, we assume that the spacing of cracks that determine the thickness of the collapsed sections is similar within each site, and therefore, the collapsed section that formed the "Farther Pile" was about 8 meters thick, similar to the "Closer Pile" thickness. The time interval between the "Farther Pile" and the "Closer Pile" is 11.4±2.2 ky. This time interval implies a cliff retreat rate of 0.7±0.1 m ky¹.

In the third site, we can distinguish two complete cycles of repeated rock falls. One is between the time of collapse of boulder TMN-7 and the time of collapse of boulder TMN-8. The distance between boulders TMN-7 and TMN-8 is 6 meters and the age difference is 17.1±6.9 ky implying a cliff retreat rate of 0.4±0.16 m ky^-1. The other complete cycle is between the time of collapse of boulder TMN-8 and the present. We assume that boulder TMN-9 (which was not dated) has fallen only recently and that the distance it has fallen from the cliff (4 meters) represents the typical distance that boulders fall from the source cliff. Thus, the age of boulder TMN-8 (13.9±4.8 ka) and the distance from its source cliff (the recent 14 meters minus 4 meters of original distance) imply a cliff retreat rate of 0.7±0.2 m ky^-1. The weathered state of the boulders in the third site compared to the fresh boulders in the Upper and Lower Valleys and the existence of only single boulders in the third site as opposed to boulder piles in the other site can be explained by rocks less resistant to chemical weathering.

Cliff retreat rates determined from the rock falls in Timna range between 0.14 and 2 m ky^-1. A more constrained range of 0.4 to 0.7 m ky^-1 is calculated from the complete collapse cycles. Similar cliff retreat rates that range between 0.1 and 0.85 m ky^-1 were calculated in arid environments, [57–60].

In Timna, rates of cliff retreat are faster by an order of magnitude than the rate that the upper surfaces are lowered by weathering (determined from samples Timna 20 and Timna 21 - 24.3±3.6 mm ky^-1). Even though cliffs retreat through the episodic process of rock fall events while the upper surface weather continuously, over time the difference in rates maintains the sharp morphology of Timna and vertical slopes (cliffs) do not roll back to become more gentle and rounded slopes. Cliff retreat obviously initiated when voids large enough opened and enabled the collapse process to begin. This space opening was probably first accomplished through incision of stream channels through the sandstone following the development of the Arava Valley topographical base level. Assuming the range of cliff retreat rates calculated in this study, the 30 m wide Lower Valley opened approximately 40- 70 ka. This range of ages determines the time at which the upstream-moving signal of the lowering base level of the DSF reached the Lower Valley site.

Conclusions

from Matmon et al. (2005:815-816)

The combined ages from the three investigated sites suggest that the boulders were deposited in three events at 31 ka, 15 ka, and 4 ka. The most likely mechanism for boulder formation in this tectonically active hyperarid region is by tectonically induced ground motion. Paleoseismic studies in the area suggest a recurrence interval of 1000–2000 yr for earthquakes GT M6. The cosmogenic and OSL age dating of the boulders suggests that boulder-formation events do not occur as frequently as GT 6 M earthquakes occur in this area Therefore, it is apparent that boulder falls in the study area do not represent a full seismic record. We suggest that boulder formation events occur during large earthquakes only after blocks of rock are sufficiently separated from bedrock by large and well developed cracks. These cracks allow the detachment of the boulders when the next earthquake occurs.

Cliff retreat rates determined from the rock falls in Timna range between 0.14 and 2 m ky¹. A more constrained range of 0.4 to 0.7 m ky¹ is calculated from the complete collapse cycles. These retreat rates are similar to those calculated in other arid regions. Field observations suggest that the retreat occurs mainly during rock fall events and that continuous weathering of the rock faces is less important. Since cliff retreat rates in Timna are an order of magnitude larger than the calculated rate of surface lowering, as calculated in the Lower Valley site, the sharp and imposing landscape characteristic to this region is preserved.

We described the spatial and temporal relations between boulders and their original in-situ locations on sandstone bedrock cliffs. In each sampled pair, a boulder and its matching cliff face yielded significantly different ¹⁰Be concentration. A 10–20% difference in ¹⁰Be concentration between boulders and their matching cliff faces imply a spatial variability in cosmogenic nuclide production rates. Differences in cosmogenic nuclide production rates arise from differences in topographic shielding and boulder geometry. The variability in shielding factors affects each boulder and its matching cliff face differently after detachment. Calculated mean ages differed by 14-32%. However, the uncertainties associated with the age calculations prevent us from distinguishing between boulder and matching cliff ages. Age calculations considered the pre-detachment accumulation of ¹⁰Be produced by deep penetrating muons. The agreement between the ¹⁰Be and ¹⁴C model ages in sample Timna 2 implies that the estimated pre-detachment ¹⁰Be concentration was correct.

In the Lower Valley, field relations between the dated boulders and sediments that accumulated upstream behind them, dated using OSL, helped to constrain the formation age of the "Closer Pile", and suggested that the older exposure age of the cliff (3.7±0.8 ka) is a more reasonable age for the boulder detachment event. In case of complicated morphological setting it is useful to combine external constraints such as field relations or other dating methods to evaluate the reliability of cosmogenic exposure ages.

Matmon et al. (2009)

Introduction

from Matmon et al. (2009:178)

In this study we explored the spatial and temporal relations between boulders and their original in situ locations on sandstone bedrock cliffs. This was accomplished by combining field observations with dating methods using cosmogenic isotopes (¹⁰Be and ¹⁴C ) and optically stimulated luminescence (OSL). Our conclusions bear both on the landscape evolution and cliff retreat process in the hyper-arid region of Timna and on the methodology of estimating exposure ages using cosmogenic isotopes.

Geological and morphological setting

from Matmon et al. (2009:178-179)

Timna is located at the western margin of the southern Arava Valley, a hyper-arid region extending from the Dead Sea basin to the Gulf of Aqaba (Fig. 1). Timna is an uplifted core of late Precambrian igneous rocks (Druckman et al., 1993) overlain by Cambrian and lower Cretaceous clastic sediments. It is a bowl-shaped valley, about 8 km in diameter, and is surrounded by cliffs of upper Cretaceous limestone and dolomite rising up to 600 m above the valley floor. Landscape development in the Timna area is controlled by the tectonic activity along the southern section of the DSF (Hannan Ginat, personal communication). Erosional truncation of the upper Cretaceous carbonate cap rocks during the Oligocene and Miocene exposed the more erodable lower Cretaceous sandstone. As the southern Arava Valley developed into a topographic base level, the sandstone was eroded and transported towards that base level, the Timna Valley was incised, and the Precambrian basement and Cambrian sediments were exposed at the base of the Timna Valley. The Cambrian sandstone in Timna reaches a thickness of about 100 m (Druckman et al., 1993) and forms cliffs that dominate the landscape. Within the Cambrian sandstone terrain, erosion is dominated both by detachment of massive boulders from the sandstone cliffs and by slow weathering of the cliff faces, as evident from 0.1 m-scale cavities (tafoni). Generally, boulders detached from sandstone cliffs are found in piles, and are usually coated with varnish. In many locations, the cliff face and the faces on the boulders match perfectly both in detail and overall morphology. Some of these boulders can be accurately traced to their original position on the cliff from which they were detached (Fig. 2).

We concentrated on three boulder sites in the Timna area (Figs. 1 and 3). The first, the Lower Valley, contains two boulder piles (one close to the source sandstone cliff face, which we call the “Closer Pile”, and one farther away, the “Farther Pile”) detached from a cliff that rises 10-14 m above the valley floor. The boulders at this site were deposited at the mouth of the Lower Valley (30X60 m), creating a dam, and causing upstream accumulation of sandy sediments on the valley floor. The second site, the Upper Valley, contains several boulder piles. We concentrated on a single boulder pile detached from a nearby 10 m cliff. At both locations boulders are not significantly affected by erosion and perfectly match the cliff faces (Fig. 2). Unlike the first two sites that contain unweathered boulder piles, the third site we examined contains three weathered boulders that are located at distances of 20, 14, and 4 meters from the source cliff.

Results

from Matmon et al. (2009:179)

We recognize three discrete rock fall events, at 31 ka, 15ka, and 4ka (Figs. 4 and 5). In this hyper arid region the most plausible triggering mechanism for rock fall events is strong ground acceleration caused by earthquakes generated by the nearby Dead Sea fault (DSF). Our record, however, under represents the regional earthquake record implying that ongoing development of detachment cracks prior to the triggering event is slower than the earthquake cycle.

Cliff retreat rates calculated using the timing of rock fall events and estimated thickness of rock removed in each event range between 0.14 m ky^-1 and 2 m ky^-1 (Fig. 8) When only full cycles are considered, we derive a more realistic range of 0.4 m ky^-1 to 0.7 m ky^-1. These rates are an order of magnitude faster than the calculated rate of surface lowering in the area. We conclude that sandstone cliffs at Timna retreat through episodic rock fall events that preserve the sharp, imposing, landscape characteristic to this region and that ongoing weathering of the cliff faces is minor.

A 10%-20% difference in the ¹⁰Be concentrations in samples from matching boulder and cliff faces that have an identical exposure histories and are located only a few meters apart, indicate that cosmogenic nuclide production rates are sensitive to shielding and vary spatially over short distances. However, the uncertainties associated with age calculations for boulder and matching cliff face pairs yielded ages that are similar within 1σ. The use of external constraints, in the form of field relations and OSL dating helped to establish each pair’s age. The agreement between calculated ¹⁴C and ¹⁰Be ages indicates that the accumulation of ¹⁰Be at depth by the capture of slow deep-penetrating muons was properly accounted for in the study.

Rock falls and earthquakes

from Matmon et al. (2009:180-181)

Many processes, including tectonic, climatic, and environmental factors can cause rockslides (Wieczorek et al., 1996). Many of these factors can be eliminated in the case of rockslides in Timna. Snowmelt, freeze and thaw effects, ground water seepage, and tree root wedging can be ruled out due to the hot and hyper-arid conditions in the area. Although rain storms and the resulting expansion of clay and salt particles in cracks is a plausible mechanism, we would expect boulders to detach one at a time rather than in groups that form large piles as is the case in Timna. Furthermore, the frequency of clay and salt wetting events is not high enough to generate proper stress in the fractures to allow boulder release. Several observations, mainly the large size of the boulders (many of the boulders have at least one dimension longer then 5 meters) and the agreement in ages of boulder piles in the sampling locations suggest that each pile of boulders was detached from its source cliff in a single event. The proximity of Timna to the DSF, where >M6 earthquakes are common, suggests that ground shaking due to seismic events is the most likely cause for the Timna boulder slides. Rock falls are sensitive recorders of strong ground motion resulting from earthquakes (Bull and Menges, 1977). Synchronous rock falls may indicate the occurrence of past earthquakes and rock fall timing may constrain earthquake recurrence intervals and magnitude. The relation between seismic events and the formation of boulders in rock falls is well established. A worldwide correlation between landslide size and distribution and variables such as earthquake magnitude and the specific ground-motion characteristics was determined by Keefer, (1984). A coseismic lichenometry model was developed in New Zealand following the discovery that lichens growing on rocky hill slopes recorded synchronous pulses of rock falls generated by historical earthquakes. The lichenometry model was used to date boulders and rock falls associated with earthquakes (Bull, 1996a,b; Bull et al., 1994; Kong, 1994; Smirnova and Nikonov, 1990).

Current measurements in the southern Arava Valley along the DSF system show no seismic activity (Shapira, 1997). However, historical evidence documents several large seismic events (Ambraseys, 1994; Amiran, 1994; Fig. 6). Paleoseismic studies in the southern Arava Valley suggest that late Pleistocene earthquakes ranged in magnitude between 6.7 and 7.1 and the average recurrence interval was 2.8±0.7 ky (Amit et al., 2002). These studies indicate that Holocene earthquakes were more frequent, with an average recurrence interval of 1.2±0.3 ky, but with smaller magnitudes that ranged between M5.9 and M6.7. Several studies in the northern Arava Valley also suggest frequent Holocene and late Pleistocene seismic activity (Amit et al., 1996; Ambraseys, 1994; Ken-Tor et al., 2001; Migowski et al., 2004; Enzel et al., 1996; Gluck et al., 1999). The time interval between the three boulder forming events recorded in this study is much longer than the recurrence interval of >M6 earthquakes in the region. However, this discrepancy does not rule out seismic motion as the mechanism for the boulder formation. We suggest that the interval between boulder forming events represents the time that is necessary for fractures to develop to the point of minimum friction between boulders and bedrock (Fig. 7). During this time, earthquakes occur and gradually enhance the opening of fractures that surround boulders. Once these fractures are sufficiently developed, the next major earthquake releases the boulders. The spacing between the fractures determines the thickness of collapsed wall during each rock fall event and the level of sandstone lithification determines the resistance of the fresh exposed rock to weathering. Continued study of additional rock piles at Timna may help better constrain the temporal frequency of rock fall events, and improve the correlation between rock falls and earthquakes.

Conclusions

from Matmon et al. (2009:181)

The combined ages of boulders from the three investigated sites suggest that they were deposited in three events at 31 ka, 15 ka, and 4 ka. The most likely mechanism for boulder formation in this tectonically active hyper-arid region is by tectonically induced ground motion. Paleoseismic studies in the area suggest a recurrence interval of 1000-2000 yr for earthquakes >M6. The cosmogenic and OSL age dating of the boulders suggests that boulder-formation events do not occur as frequently as >6M earthquakes occur in this area Therefore, it is apparent that boulder falls in the study area do not represent a full seismic record. We suggest that boulder formation events occur during earthquakes only after blocks of rock are sufficiently separated from bedrock by large and well developed cracks. These cracks allow the detachment of the boulders when the next earthquake occurs.

Cliff retreat rates determined from the rock falls in Timna range between 0.14 and 2 m ky^-1. A more constrained range of 0.4 to 0.7 m ky^-1 is calculated from the complete collapse cycles. These retreat rates are similar to those calculated in other arid regions. Field observations suggest that the retreat occurs mainly during rock fall events and that continuous weathering of the rock faces is less important. Since cliff retreat rates in Timna are an order of magnitude larger than the calculated rate of surface lowering, as calculated in the Lower Valley site, the sharp and imposing landscape characteristic to this region is preserved.

Agnon (2014)

from Agnon (2014:232)

A novel approach to earthquake driven gravitational collapse on slopes is given by Matmon et al. (2005). They use cosmogenic dating together with OSL for dating collapse of several meter large sandstone boulders in Timna (Fig. 8.8). They identify three distinct events around 3–4.5 ka, 15 ± 1 ka, and 31 ± 5 ka respectively. Notably the oldest event recorded is associated with a ~5 m boulder displaced 20 m horizontally from its source cliff, with merely 2 m vertical drop. The boulder is found due north of its north-facing scar, suggesting a significant role of horizontal acceleration parallel to the slip vector. Matmon et al. (2005) point out that the cycle of boulder shedding off the cliff may be limited by either of the steps: undercutting and ground shaking. The apparent cycle of 12–15 ka may be determined by the rate of undercutting after a large earthquake. On the other hand, a similar period (yet with a different phase) seems to emerge from Kagan et al. (2005) cave deposits (discussed in the following section). Further research is required for ruling between these proposed mechanisms for determining the period of boulder shedding events.

Master Seismic Events Table

References

Articles and Books

Agnon, Amotz, Amit, Rivka, Hough, Susan, and Michetti, Alessandro Maria Michetti (2011) Introduction: The Dead Sea Rift as a natural laboratory for neotectonics and paleoseismology, Israel Journal of Earth Sciences Vol. 58

Agnon, Amotz (2014). Pre-Instrumental Earthquakes Along the Dead Sea Rift: 207-261 in Dead Sea Transform Fault System: Reviews, Springer Netherlands, Dordrecht.

Matmon, A., et al. (2005). "Landscape development in an hyperarid sandstone environment along the margins of the Dead Sea fault: Implications from dated rock falls." Earth and Planetary Science Letters 240(3–4): 803-817.^A

Matmon, A. et al. (2006) Corrigendum to “Landscape development in an hyperarid sandstone environment along the margins of the Dead Sea fault: Implications from dated rock falls” [Earth Planet. Sci. Lett. 240 (2005) 803–817]

Matmon et al (2009) Landscape development in an hyper arid sandstone environment along the margins of the Dead Sea fault: implications from dated rock falls in The Dead Sea Rift as a natural laboratory for Earthquake Behavior Prehistorical Historical and Recent Seismicity INQUA field guide DSW 2009 part 1, Arava Valley, pp. 178-191

Rinat, Y., et al. (2014). "Holocene rockfalls in the southern Negev Desert, Israel and their relation to Dead Sea fault earthquakes." Quaternary Research 81.

Rinat, Y., et al. (2014). "Holocene rockfalls in the southern Negev Desert, Israel and their relation to Dead Sea fault earthquakes." Quaternary Research 81. SUPPLEMENTARY MATERIAL

Footnotes

Sometimes this articles is listed, cited, or titled as :

Matmon, A., et al. (2005). "Seismic ground motion and boulder detachment along the margins of the Dead Sea fault, southern Israel." Earth and Planetary Science Letters 240: 803-817.

or

Matmon, A., et al. (2005). Lessons to exposure age dating from constraining the time of earthquake induced rockfalls along the margins of the Dead Sea fault system, southern Israel. Earth Planet Sci Lett 240: 803-817.