Palaeoseismic trench study of the Jordan Gorge Fault

Partners:

Thomas K. Rockwell, San Diego State University, California

Amotz Agnon, The Hebrew University, Jerusalem

Ariel Heimann, The Geological Survey of Israel, Jerusalem

Uri Frieslander, The Geophysical Institute of Israel, Lod

Publication in review

Abstract

Three-dimensional excavations of buried stream channels that have been displaced by the Jordan Fault, the primary strand of the Dead Sea fault zone in northern Israel, demonstrate that late Holocene slip has been primarily strike-slip at a minimum rate of 3 mm/yr. The palaeoseismic study was carried out in the Bet-Zayda Valley, the delta of the Jordan River at the north shore of the Sea of Galilee. The site was chosen where five indicators for the fault have been identified: 1) It is aligned with the previously-confirmed fault of the Jordan Gorge, 2) A lineament seen in Landsat images and in airphotos, 3) A major fault identified in deep seismic reflection, 4) Offsets of shallow Plio-Pleistocene reflectors are seen in a high-resolution seismic image, and 5) A north-striking scarp with up to one meter vertical expression crosses the flat valley. One group of trench excavations was located where a small stream crosses the scarp. The active stream, which is incised into the scarp is not offset by the fault. However two palaeo channels about 2 m below the surface were found to be offset sinistrally 2.7 m by the fault and two younger nested channels are offset by about 0.5 m. Based on radiocarbon dates we attribute the last 0.5 m rupture to the earthquake of October 30, 1759. The older offset of 2.2 m most probably occurred in the earthquakes of May 20, 1202. These two events correlate with the findings at Ateret, about 15 kilometers north of Bet-Zayda, where the 1202 earthquake produced 1.6 m of lateral displacement in the E-W-striking defence walls of a Crusader castle, and an Ottoman mosque was offset 0.5 m in the earthquake of 1759. In the second group of trenches some 60 m further south we found another offset channel. Its northern margin is displaced 15 m sinistrally whereas the southern margin shows only 9 m of sinistral offset. The dip slip component is 1.2 m, west side down. The different amounts of margin offset can be explained by erosion of the southern margin during the first 6 m of displacement. Additional slip of 9 m that accrued after the stream had been abandoned and buried by a 2-m-thick lacustrine clay layer. Radiocarbon dates on organic residue provide the age control which indicates that the 15 m of slip has accrued over the past 5 kyrs, yielding a short-term slip rate of 3 mm/yr for the late Holocene. It is possible that our study covers only part of the fault zone, hence we regard this average to be a minimum rate for the DST. Based on other palaeoseismic studies the best estimate for Quaternary slip rate is 4±1 mm/yr.

The topography and the major faults of the Dead Sea Transform system in northern Israel (after Bartov 1979 and Ben-Avraham et al. 1996). Shaded relief by Hall (1994).

Inset: Tectonic plates in the Middle East.

An airphoto of the study area. back to top

East West

A straight N-S-trending escarpment, up to 1 m high, crosses the otherwise extremely flat floodplain north of the Sea of Galilee (Kinneret). Ezra Zilberman stands on the upper block on the left (east) and Amotz Agnon on the downthrown side. back to top

Ariel Heimann (left) and Shlomo Ashkennazy operate a water pump.

Meir Abelson and Yariv Hamiel cover the trenches before covering. In the following trenching season a year later we could return to exactly the same trench walls.

Photos of fault planes.

Reconstruction of stream channels. In the southern group of trenches the offset of the northern margin of the channel, which is dated at 5 ka, is 15 m. Its southern margin is displaced only 9 m. We explain the difference by erosion of the opposing corner. Two channels in the northern group of trenches, CH2 and CH3, which postdate C14 dated layers of up to about 12^th century AD and predate layers of 13^th century AD, are offset 2.7 m. Two younger channels that postdate the 15^th century AD are offset 0.5 m. We attribute these offsets to the May 20, 1202 and the October 30, 1759 earthquakes, the same earthquakes that were identified at Ateret.