Paleoseismic Records of the Dead Sea Basin Reveal Climatic Modulation of Seismicity Along the Continental Transform Fault
Summary
The longest detailed paleoseismic record acquired along the Dead Sea Fault (DSF) provides an excellent opportunity to study long-term temporal variability (clustering) of earthquake occurrence along a continental transform fault. The record was derived from borehole and outcrop observations of seismites, which are sediments deformed by earthquake shaking in the lakes of the Dead Sea Basin over the past 220 ka. The seismite sequences provide the most complete temporal record of earthquake occurrence in a single location. Preliminary analysis of the 220 ka record reveals a significant high correlation between long-term earthquake occurrence and lake levels, suggesting a climatic impact on earthquake occurrence frequency. The project investigates how lake level changes can impact the paleo-earthquake record as derived from the Dead Sea seismite sequence.
Background
The Dead Sea Fault is a major tectonic boundary extending over 1,000 kilometers, separating the African and Arabian plates. Over the past 20 million years, it has accommodated approximately 100 kilometers of left-lateral (northward) movement of the Arabian plate relative to the African plate. This displacement is supported by multiple lines of evidence, including regional plate tectonic reconstructions, geological formations, geodetic measurements, and both historical and prehistoric seismic records. Contemporary geodetic data indicate that the fault continues to be active, with slip rates ranging between 4.2 and 5.8 millimeters per year, consistent with long-term geological observations.
An en-echelon arrangement of the fault segments results in a series of pull-apart grabens, which subside faster than the sedimentary fill, thereby forming deep morphological basins, e.g., the Sea of Galilee and the Dead Sea. The basins hosted a series of lakes that grew during glacial periods and shrank during interglacials.
The 220 ka seismite records in the Dead Sea Basin
During the Quaternary period, a series of lakes repeatedly formed and disappeared in the pull-apart basins along the Dead Sea Fault. These ancient lake sediments, found throughout the region, reflect global climate changes, as lake levels rose during ice ages when ocean levels dropped. The fine layering of these lake deposits provides a detailed record of past seismic activity. Distinct features, such as folded layers and breccia formed underwater, reveal strong ground shaking from earthquakes. These features often appear next to faults that were active during sediment deposition and have been linked to historical earthquakes over the past 3,000 years. The most comprehensive record comes from deep sediment cores collected through the Dead Sea drilling project, covering the last 220,000 years, making it the longest continuous earthquake record in the world.
Figure 2. Examples of seismites in outcrops (left) and in ICDP Core 5017-1 (right).
Project Objectives
1.Characterize the 220,000-Year Seismite Record
The project aims to analyze the extensive seismite record obtained from the Dead Sea Basin to explore its correlation with long-term climate variability. Preliminary findings suggest a strong relationship between seismite frequency and lake-level fluctuations, which reflect broader climatic changes. This component focuses on aligning the temporal distribution of seismites with available regional climate proxies.
2.Assess Uncertainty in Paleo-Earthquake Detection
The current long-term earthquake record is derived primarily from a single deep core and several outcrop exposures, leading to uncertainties in event detection. This component evaluates the reliability of using a single core to represent basin-wide seismic activity, through detailed comparisons with multiple outcrop sections to identify the spatial and temporal variability in seismite preservation.
3.Estimate Sediment Availability for Turbidite Formation
The study will assess how lake level and topography influence the availability of sediments to form seismically triggered turbidites. Using bathymetric and topographic data, combined with rheological measurements from modern analogs, we will model slope stability and the probability of turbidite deposition under different hydrological conditions.
4.Evaluate Lake-Level Effects on Sediment Rheology and Deformation
Changing lake levels, driven by climatic factors, alter the physical properties of lakebed sediments. This component will use numerical modeling to investigate how sediment rheology, slope angles, and hydrostatic pressures influence deformation styles during seismic shaking events.
5.Quantify the Impact of Lake-Level Changes on Earthquake Frequency
Variations in lake level can affect fault stress through hydrological loading and pore pressure changes. Using 3-D boundary element and finite element models, we will quantify Coulomb Failure Stress changes across both strike-slip and normal faults to evaluate how seismicity rates vary with lake-level fluctuations over timescales of thousands to hundreds of thousands of years.
6.Support Student and Postdoctoral Training
The project includes a strong educational component, providing research opportunities for an undergraduate student and a postdoctoral researcher at Florida International University, a minority-serving institution. Additionally, it supports a graduate student and a postdoc at Tel Aviv University through U.S.–Israel collaborative funding.
Preliminary results
Characterization and interpretation of the 220 ka record
The 220 ka record presents a large temporal variability of seismite occurrence, in terms of periods with a high number of seismite occurrences versus periods with a low number of occurrences, aka ‘temporal clustering’ (Fig. 3a). This general variability occurs in three orders of magnitude, suggesting that the temporal behavior of seismite occurrence has fractal characteristics. The occurrence variability is reflected in the total number of events per period, which also shows large variability in three orders of magnitude (Figure 3b).
Spectral analysis of the binned seismite occurrences reveals four dominant periodicities of 14, 20, 55, and 110 ka. The high periodicities of 55 and 110 ka are derivative of the time series length and are apparent in the histogram of the 10 ka record. Conducting spectral analyses with long sub-sets of the entire sequence and various bin sizes revealed variations in the actual dominant periodicities. But the general pattern of the two groups of periodicities remains. The two groups are of long periodicities of 60-110 ka and shorter periodicities of 14-20 ka.
Figure 3. (a) Seismite occurrence plotted at 3 time scales: 220 ka, 22 ka, and 2.2 ka. The record suggests a fractal-like occurrence pattern. (b) Temporal distribution of seismite occurrence at 3-time scales: 10 ka, 1 ka, and 0.1 ka. The record show high temporal variability in the number of seismites per period. Spectral analysis of the binned seismite occurrences reveals four dominant periodicities of 14, 20, 55, and 110 ka.
The longest detailed paleoseismic record (220 ka) obtained from a core analysis of the paleo-Dead Sea sediments reveals high temporal earthquake occurrence variability. A preliminary statistical analysis of the record revealed large fluctuations in earthquake occurrence frequency, in which the number of earthquakes per 104 years varied in the range of 10-110 (Figure 4a). Interestingly, periods with high earthquake occurrence occurred during high stands of the Dead Sea, and periods with low earthquake occurrence occurred during low Dead Sea levels (Figure 4b). A correlation analysis between the earthquake occurrence frequency and Dead Sea level revealed a strong correlation (Figure 1c; R2 = 0.67), suggesting a possible contribution of climatic processes to the observed earthquake occurrence fluctuations. We present the results of this preliminary analysis in the research overview because the proposed project is aimed at understanding this very interesting phenomenon.
Figure 4. (a) Earthquake occurrence frequency of the past 220 ka. The horizontal solid line indicates the expected number of events per 10 ka due to strike-slip faulting along the DSF. The dashed lines mark the uncertainties in our estimate. (b) Dead Sea level derived from Guillerm et al. (2023). The horizontal solid red lines mark the 10 ka mean level of the record. (c) Scatter plot of earthquake frequency versus Dead Sea level showing a significant, high positive correlation (R2 0.67 and p-value < 0.001).
Level changes impact on the sediment rheology and deformation style
To investigate the formation of seismites, researchers explored the role of Kelvin-Helmholtz Instability (KHI), which occurs when velocity shear develops at the interface between fluid layers of different densities. This instability produces a range of deformation structures, evolving from linear waves to asymmetric billows, coherent vortices, and ultimately turbulent mixing. Initial numerical models assumed Newtonian rheology for both the water column and underlying saturated sediments but failed to capture the full complexity of observed deformation (Figure 5). More advanced simulations incorporate non-Newtonian, viscoplastic behavior using a shear-thinning model to better represent the mechanical properties of consolidated lacustrine muds. The model domain includes three stratified layers—saline water overlying sediment layers with increasing density and viscosity—and applies the incompressible Navier-Stokes equations under shallow-water assumptions, with appropriate boundary conditions. Preliminary results reveal distinct differences in flow dynamics under varying rheological conditions, and the model will now be used to assess the sensitivity of deformation structures to changes in water depth, with the aim of reconstructing paleo-lake levels.
Figure 5. Numerical simulation of a sedimentary layers deformation in response to seismic shaking. Snapshot at 1.3s-on. (Top) Newtonian fluid for all layers; (bottom) Viscoplastic fluids for mid and lower layers. Ground acceleration: 10 m/s^2; saline water (green) layer thickness: 1.9 m; mid (red) and bottom (dark red) layer thickness: 0.05 m; mid layer density: 1600 kg/m^3; bottom layer density: 1750 kg/m^3; mid layer kinematic viscosity: 0.3 Pas; bottom layer kinematic viscosity: 3 Pas; mid and bottom layer yield stress: 0.1 Pa; mid and bottom layer flow index: 0.2 (Thinning).
Mechanical modeling of earthquake occurrence at various lake levels
To examine how high lake levels in the Dead Sea might have influenced the frequency and magnitude of past earthquakes, a 3D boundary element model was developed using a simplified pull-apart basin geometry consistent with the left-lateral motion of the Dead Sea Fault. The model was driven by a strain tensor derived from horizontal GPS velocities, representing long-term tectonic loading over a 1,000-year interseismic period. Coulomb Failure Stress (ΔCFS) was calculated to assess how stress builds up on faults. The model incorporated both surface loading from varying lake water levels (20 m and 250 m) and pore pressure changes within the crust. Pore pressure was modeled based on undrained fluid conditions with depth-dependent permeability and viscosity, leading to pressure increases of several tens of kilopascals at depths around 5 km. The results indicate that horizontal shear along the transform faults generates ΔCFS of several tens of kilopascals over interseismic timescales. At high lake levels, the added weight of water significantly increases stress on high-angle normal faults, with ΔCFS exceeding 100 kPa, potentially triggering earthquakes. When pore pressure effects are also considered, both the transform and normal faults show elevated ΔCFS values above 100 kPa at seismogenic depths (<7.5 km), suggesting that elevated lake levels can promote earthquake occurrence on multiple fault systems. Importantly, this stress accumulation at depth allows for larger rupture areas along the transform fault, implying a greater potential for high-magnitude earthquakes under high lake-level conditions.
Figure 6. Numerical modeling of stress changes on the primary left-lateral strike-slip and secondary normal fault systems bounding a simplified pull-apart basin. The color scale denote Coulomb Failure Stress change (ΔCFS) after 1000 years of strain accumulation for a 1,000-year interseismic period for (a) low lake level of 20 m, (b) hydrostatic pressure (hydrologic load) of high lake level of 250 m, and (c) hydrostatic pressure and change in pore pressure due to fluid flux for high lake level of 250 m. The color scale is the same for all panels.