T41E-01 INVITED
The Use of Paleoseismic Data of Strong Earthquakes for Seismic Hazard Evaluations of the Dead Sea Transform.
Synthesis of paleoseismic data along the Dead Sea Transform shows that different segments have a specific pattern of time distribution of strong earthquake. It was found that during the last 20 ky the segments of the Arava and the Jordan valley have recurrence intervals of strong events, M > 6, of about 1 ky, while in the Hula valley the recurrence interval is 350 yr and in the Dead Sea segment it is 3-5 ky. However, despite differences between the segments some similarities can be shown, especially in the long time frame. At least two segments, the Arava and the Dead Sea, show a similar change in earthquake pattern over time. In both segments the probability of occurrence of strong earthquakes decreased gradually with time over the last 60 ky. In the southern Arava the magnitude range of earthquakes that occurred between 80 ka and 20 ka is M6.7 - M7 with average recurrence intervals of about 3 ky, whereas the magnitude range during the last 20 ky is M5.9 - M6.7 with average recurrence intervals of about 1 ky. Although the level of the recent seismic activity in these two segments is different, a similar trend of decrease in earthquake magnitudes with time was observed in the Dead Sea segment. Contrary to the paleoseismic evidences it is possible that the current seismic activity in the southern Arava, may point to a seismic gap. This may rely on the following observations: a. The temporal decrease in paleo- earthquake magnitude; b. The observation that the last large event occurred in A.D. 1068; c. The low rate of microseismicity recorded in the instrumental period since the early 1980's. In addition, these paleoseismological studies show that over the last 100 ka no evidence was found for earthquakes larger than M 7.5. It appears that there is an upper limit to the magnitude of the events that can be produced by the DST. It is suggested that this magnitude limit is an inherent characteristic of the DST which is controlled by the structure and the dimension of its segments.
T41E-02 INVITED
Paucity of Historical Earthquakes South of the Dead Sea: Is it Real?
The long cultural record of Dead Sea Transform (DST) earthquakes provides a unique opportunity for studying the patterns of seismicity over multiple seismic cycles, especially if rupture and/or epicentral areas of historical earthquakes can be identified. It has long been noted that the modern rate of seismic activity in the Wadi ‘Arabah (Arava) south of the Dead Sea is lower than other segments of the DST and few historical earthquakes have been located south of the Dead Sea. Does the tendency to assign most epicentral locations north of the Dead Sea reflect a bias toward higher ancient settlement patterns? Seismites studied in Israel suggest a millennial-scale clustering of seismicity with a paucity of activity during the 6th-11 centuries. However, comparison of the timing of fault rupture constrained by paleoseismic trenching, archaeoseismic excavations, and radiocarbon and archaeological artifact dating from sites on the northern Wadi ‘Arabah fault (Qasr Tilah area) and in Aqaba in Jordan suggests a temporal correlation of earthquake activity along the full length of the DST south of the Dead Sea during this quiescent period. Our data indicate a major period of seismic activity during the late 6th to 8th centuries. Furthermore, faulting and architectural collapse during the 2nd and 4th centuries indicate other periods of increased earthquake activity. Evidence for 2-3 additional ruptures in the 9th-16th centuries that are attested to by historical accounts, while evident in the geologic faulting record are poorly constrained by independent age control. At least one of these events is likely the 1068 earthquake. Correlation of earthquakes from the north and south ‘Arabah indicate either very long fault rupture (about 160 km) or multiple ruptures on adjacent fault segments that are closely spaced in time. In Aqaba, multiple ruptures during these earthquake intervals may indicate both foreshock and main events similar to the seismic sequence preceding the 1995 Nuweiba earthquake. Greater coseismic slip in the pre-11th century earthquakes compared to the last millennium suggests large potential strain accumulation and a long elapsed time since the last major earthquakes south of the Dead Sea with important implications for the seismic hazard of the region. Furthermore, these data suggest that historical earthquake catalogues are incomplete with regard to some of the earthquakes that have affected southern Jordan.
T41E-03
THE TECTONIC GEOMORPHOLOGY AND THE ARCHEOSEISMICITY OF THE DEAD SEA TRANSFORM IN JORDAN VALLEY
The Dead Sea transform (DST) extends 1000 km from the Sinai triple junction in the south to the Tauros- Zagros collision zone in Turkey in the north. In Jordan, the DST consists of three morphotectonic elements; the Wadi Araba in the south, the Dead Sea basin in the middle and the Jordan Valley in the north. The Dead Sea is a pull- apart basin that formed due to the overlap of the Wadi Araba fault (WAF) and the Jordan Valley fault (JVF). The movement along the transform is active as indicated from both the geomorphological features and from the seismic activity. The DST is a major left lateral strike slip fault that accommodates the relative motion of the Arabian plate to the east and the Sinai plate to the west, where 107 km of cumulative left lateral offset has occurred over the last 18 million years. Based on this offset, the accumulated slip rate is estimated to be 5-10 mm/yr. Based on aerial photographic analysis of the DST and earthquake catalogue information, it is suggested that the present day slip rate has been slower (1.5-3.5 mm/yr) when compared with the Pleistocene rates. Recent work on offset alluvial fan surfaces and drainage along the northern Wadi Araba fault indicates a slip rate of 4.7 mm/yr (Niemi et al., 2000) and 4 mm/yr (Klinger, 2000). In the Jordan Valley fault a slip rate of 7 mm/yr in the last 13000 years was estimated based on aerial photograph and satellite image interpretation (Al-Taj, 2000). Active strike slip faults display distinct morphological features along its trace. The DST in Jordan Valley has a series of morphotectonic features, such as pressure ridges and sag ponds. These features are formed in the place of fault steps or bends (Keller and Pinter, 1996). Fault scarps are formed along most of the trace indicating a dip slip component of displacement. Historical, archeological and paleoseismic data are combined from two trench sites to build a unique composite catalogue of large past earthquakes. On that basis, evidence for surface rupture during the AD 749 and AD 1033 earthquakes was shown. Overall, 8 surface-rupturing events for the last 14 kyr were identified. A temporal analysis displays clusters of seismicity as well as quiescence periods as well as a 600- to 1000-yr-long recurrence interval for large earthquakes in the last 14 kyr. http://www.hu.edu.jo
T41E-04
A Tale of Two Cataclysmic Earthquakes: 39 and 52 kyr BP, Dead Sea Transform, Israel; a Multi-archive Study
We have documented earthquake histories in four lacustrine sites and a cave in the Dead Sea Transform region in central Israel. The lacustrine Lake Lisan (last Glacial paleo-Dead Sea) sites include: Massada Plain (M1b), Perazim (PZ1), Tovlan (NT), and Tamar (TM). They are up to 110 kms apart, along the Dead Sea Basin. These lacustrine sites have a variety of deformed marls (e.g. brecciated, homogenated, folded, and/or faulted). Except for the more fluvial NT site, where there is only one breccia layer, the sites show numerous (up to 29) earthquake events. Brecciated marls have been shown to be valuable earthquake markers by correlation with historical earthquakes and by their relationship to intraformational fault scarps (Agnon et al., 2006). The Soreq Cave, a carbonate cave richly decorated with speleothems, is 40 km west of the Dead Sea Basin, near the town Bet Shemesh. Earthquake damage in the cave includes collapsed stalactites and ceilings and severed stalagmites. During the last Glacial time period the cave, more distant from the Transform than the lake sites, experienced ~7 damaging events, documented by tens of dated collapses. The Soreq cave collapses have been shown to be viable earthquake markers by correlation to lacustrine documented seismic events and by absence of potential non-seismic sources of damage in the cave (Kagan et al., 2005). All the earthquake evidence, speleological and lacustrine, was rigorously dated by high resolution mass spectrometry by MC-ICP-MS at the Geological Survey of Israel. Both the Soreq cave and the Lisan sediments have been studied intensely for paleo- climate purposes in other studies. From these different and distant paleoseismic sites two events stand out. At ~39±1 ka and ~52±2 ka there is paleoseismic evidence at 5 and 3 sites, respectively. The later event, ~39±1 ka, has left evidence of brecciated marls at all four Lisan sites (with extremely thick seismites at the PZ1 site and the only breccia at the NT site) as well as five well-constrained collapses in different areas of the Soreq cave. The ~52±2 ka seismic event has left evidence at M1b (a closely spaced doublet of seismites), at PZ1 (followed by a period with numerous (8) and very thick (up to 48 cm) seismites) and at the cave site (3 collapses). We suggest that these seismic events are large-earthquakes, sufficiently strong to cause cave collapses and lake bottom brecciation in numerous and distant sites in very different kinds of sediments. We suggest that these large events are representatives of the largest magnitude earthquakes in the Dead Sea region. We extrapolate the frequency-size statistics from the instrumental data using a power-law (Gutenberg-Richter) to estimate ranges of magnitude for these cataclysmic events.
T41E-05
Slip rate of the Wadi Araba fault, southern segment of the Dead Sea fault, derived from GPS and geomorphic measurements
The Dead Sea Fault (DSF) accommodates the northward displacement of the Arabia plate relative to Sinai. Although it counts among the major strike-slip faults, no agreement has been reached yet about its slip rate, either instantaneous or longer term. Proposed values vary from 2 to 10 mm/yr. We present here an extensive study of the slip rate for the southern segment of DSF, Wadi Araba fault, based on GPS profiles and measurements of offset geomorphologic features, which ages are comprised between 10 and 350 kyr. We installed 17 campaign-style GPS sites distributed along 3 profiles perpendicular to Wadi Araba fault with far-field points up to 90 km away from the fault. The sites have been measured twice, in 1999 and in 2005, during 48h- long sessions. Measurements are complemented by data from permanent GPS stations in Israel. Using a locked fault model, we estimate the slip rate to be 4.9 ± 1.6 mm/yr over 6 years. Tests for possible creep on the fault show that if it exists it is not significant. To estimate the slip rate over longer periods of time, we targeted alluvial fans offset by the fault at 4 sites, that we mapped and sampled for 10Be cosmogenic dating. At one site, preferred offset of 46 ± 4 m of a surface dated at 11 ± 3 kyr yields a slip rate of 4.7 ± 1.9 mm/yr, in very good agreement with the present-day slip rate. At the other sites, preliminary reconstructions show offsets of ~600 m to ~5400 m with ages ranging from 50 to 350 kyr and suggest a preliminary slip rate of 4-9 mm/yr. Ongoing studies aim to better understand the history of the different lobes of the fans, from their emplacement to their abandonment, to reach a tighter constraint on the geomorphologic slip rate. Interestingly, the slip rates presented here, that have been determined along the same transects of the DSF, show good agreement despite differences of methodology (GPS and geomorphology). It suggests that discrepancies between slip rates determined at different time scales could be due to misunderstanding of fault behaviors at different time scales rather than due to methodologies.
T41E-06
Quaternary Transform Kinematics Constrained by Sequence Stratigraphy and Submerged Coastline Features: The Gulf of Aqaba
The northern head of the Gulf of Aqaba is a zone of transition between major segments of the Dead Sea Fault System (DSFS), and where this fault system crosses the evolving continental shelf. This paper is based mainly on detailed processing and interpretation of a grid of high resolution Chirp sub-bottom profiles we acquired in 2002 along the western side of this shelf down to a depth of about 120m. Our data reveal stepping seafloor morphology comprising a series of relict coastline features, primarily fossil reefs, that reached their current depth as a combined result of eustatic sea level rise and tectonic displacements. On the north slope the fossil reefs are embedded in up to about 30 m thick, predominantly single phase, retrograde sedimentary stack that evolved presumably during the most recent phase of sea level rise since the last glacial maximum, about 20 Ka ago. Large variations in the sedimentary character within this sedimentary stack bear evidence for large episodic changes in the rates of sea level rise and the ensuing environmental conditions during the Holocene. A prominent terrace-face complex, a relict coastline feature, sub-parallels the shoreline outlining at a depth of about 100 m an internal basin. This face-terrace complex accommodated on the order of 10 m down to the east vertical offsets across the surveyed area since its creation, presumably late Pleistocene to early Holocene. The trace of an active strike slip fault crossing the north slope is defined by: a. NE striking bathymetric features; b. discontinuities offsetting the sedimentary layering from just below the sea floor to the deepest layers imaged; c. a 30±10 m sinistral and 10±1 m down to the east offset of a fossil reef at a depth of about 65 m. Global sea level curves constrain the age of the 65-m reef to 11±2 Ka yielding an estimated average Holocene sinistral slip rate of 2.7±1.5 mm/yr, in the order of half the regional slip of the DSFS. We conclude that the revealed fault is a principal strand of the DSFS continuing to the south the Avrona fault in a left stepping manner. Segmentation to echelon segments, as inferred here for the Avrona fault, may be a typical pattern for the accommodation of strike slip at tips of transform basins.
T41E-07
Radon Along the Dead Sea Transform: a Proxy of Geodynamic and Geophysical Processes
Radon (Rn-222) occurring in geogas and water phases in upper crustal levels is often mentioned as a proxy of active geodynamics, with a potential as a precursor of earthquakes. So far the reported variability and complexity of its signals limits establishment of a verifiable scientific basis for this proposition. Most investigations tried to relate the temporal variations to different combinations of atmospheric, hydrologic, geophysical-geodynamic as well as geochemical drivers. An extensive set of stations, measuring radon with a resolution of <1-hour in the unsaturated zone is implemented since 1995 in a 200-km sector along the western margin of the southern DST. Several signal types comprise the measured signal – multi-year variation, a periodic Seasonal Radon (SR) signal, a multi-day (MD; 2-20 days) signal and periodic Diurnal Radon (DR) and sub-diurnal signals (SDR; several hours). Analysis in both the time and frequency domain – per individual sites and per groups of sites situated in identical arid climatic conditions - negates atmospheric influences as the primary forcing of the radon signals in the geogas system, albeit apparent similarity. A mechanical-geodynamic (subsurface) driver is suggested in cases where: a) relation can be shown between temporal and periodic features in the time series of radon and local geologic/structural situations; b) Frequencies typical of Earth tide (M2; O1) occur in the diurnal variation band; c) a statistically significant correlation is established between MD Rn signals as recorded in the Dead Sea and earthquakes that occurred within the Dead Sea rift valley several days after the start times of these anomalies; d) when a bigger amplitude of the DR signal is associated with the rising limb of a MD signal. Characterizing the fundamental statistical properties of radon time series indicates that further and different external (above surface; non-atmospheric) processes determine the periodic (SR, DR, SDR) temporal variation of subsurface radon. The main criteria are: a) cases where the amplitudes of the S1 and S2 diurnal frequencies, dominating the DR signal, vary in a regular seasonal variation pattern; b) the ratios (per 21.3 days) of co-occurring amplitudes of S1 and S2 frequencies define a linear pattern; c) Locations where radon time series are non- stationary in mean and in variance combined with a coupling between variance and mean-level; d) Large SDR signals occurring at depth (85m) in some days with a preferred bi-modal incidence within the 24-hour cycle. The radon signal in subsurface environs is driven by both internal (primarily the MD signal) and external (SR?, DR, SDR signals) driving processes. The first may be linked to mechanical geodynamics and the second may be related to an unidentified extraterrestrial influence - possibly related to solar irradiance. Both mechanisms interact with the local subsurface geological condition to yield the highly significant and complex phenomena observed in radon time series.
T41E-08
Earthquake Clustering along the Dead Sea Fault: The Influence of Strain Pattern and Geometrical Complexities on Rupture Propagation
The temporal clustering of large earthquakes is a salient characteristic of major continental faults in active zones. Interesting examples are large earthquake clusters along the East Anatolian fault (1820 – 1905), the North Anatolian fault (1912 – 1999), the Dead Sea fault (1137 – 1293), the southern San Andreas fault (1502 – 1680), the Kunlun Fault (1937 – 2001), the Tien Shan fault system (1885 – 1992) and Bulnay-Bogd fault system (1905 – 1957). Recent projects and faulting studies with paleoseismic investigations along the Dead Sea Fault (DSF) provided a wealth of field data and results on the physical characteristics of earthquake ruptures. Using individual and cumulative slip, and the rich historical seismicity catalogue and archeoseismic investigations along fault strike, I examine the length of earthquake rupture segments, the slip rate and timing of past earthquakes. The detailed mapping of rupture zones showing structural restraining bends, releasing step-overs, patch and segment boundaries, and slip distribution along strike illustrate their geometrical complexities. Taking into account the geologic and geodetic slip rate, I observe that the long-term behaviour of fault segments and/or patches determines the occurrence of seismic sequences and the location of seismic gaps. In addition to the slip deficit inferred from paleoseismic studies, two sections along the DSF shows seismic quiescence since nearly 1000 years and may be the location of near future earthquakes. In most cases, the clustering of large earthquakes migrate along fault segments and show off sequence seismic events. The mechanical coupling between off sequence distant earthquakes and laterally propagating ruptures depend mostly on the stress change at fault discontinuities and related block tectonics. The temporal clustering and multi-segment earthquakes ruptures in the past with coupling between step-overs and stress change suggests the size and probable length of future large earthquakes along major continental faults. http://apamea.u-strasbg.fr/