Work package number: WP6

Work package title: Earthquake-Induced Landslide Hazard in Marmara

WP Leader: INERIS

Objectives

The 1999 Izmit and Duzce earthquakes in northwest Turkey have revealed the Ataköy area located westwards of Istanbul as affected by very significant local site effects (Sørensen et al., 2006). Moreover, the adjacent urbanized and geologically similar area of Cekmece has been geologically and geotechnically surveyed and characterized as a concentrated landslide prone area, showing high susceptibility to both landslide and liquefaction phenomena (Duman et al., 2005).

This fast developing area includes also critical facilities such as the Atatürk international airport and several industrial plants. Latest earthquake disasters underline how important better prediction of those geohazards is for the prevention of casualties and economic losses. Eventually, offshore landslide tsunamigenic hazard triggered by strong earthquake is clearly to be considered when reflecting about the closed situation of the Marmara sea.

The aim of this work package is to improve the preparedness of those seismically induced landslide geohazards, through the using and the improvement of monitoring and observing systems in hydrogeotechnical and seismically well-constrained areas within the supersite. Two areas,on-shore and off-shore, will be studied deeply to gain knowledge and improve the capabilities to work out guidelines for a LHS “Landslide Hazard Scale”, including earthquake triggering factor. The terrestrial western part of Istanbul and a potential submarine landslide detected at the entrance of the Izmit Gulf are the two identified targets.

As regards landslide pre-disposition, pre-existing geological, geomechanical and geomorpholocial and geophysical data, including high-resolution DInSAR data, will be selected and analysed to develop better understanding and enhance capabilities to hazard assessment and susceptibility mapping, including local site effects enabling the earthquake triggering of inactive or dormant landslide. Gain is expected as regards methodology for areas to be studied in the future (Task 1).

As regards local site effects, considering the pre-existent earthquake scenarios based on the closest NAF segment in the Marmara Sea (Pulido et al., 2004), ground motion modelling showed that highest ground motions are expected in this area, obviously due to its very close vicinity to the NAF. Ground motion data currently collected will be processed and modelled to grid and map site effects and to test them versus case studies (Task 2).

Description of work

Task 1. Investigations of local instability areas – onshore and offshore – and developing of advanced susceptibility mapping

It is well known that the shelves and slopes of the Sea of Marmara are prone to landslide/tsunamigenic hazard caused by the diffuse presence of potentially instable sedimentary bodies that could slump towards the basins centre as a consequence of major earthquakes. In fact, this has occurred during several strike-slip earthquakes, with mainly horizontal displacement, that have, however, caused local although destructive tsunamis, probably due to secondary mass movements caused by the shaking (Yalciner et al., 2002). The first step towards a mitigation of landslide-derived tsunami hazard has been the mapping of all potential gravitational sliding through the use of the dense-spaced marine geophysical database available. Thus, all previous geophysical surveys will be closely examined.

The first geophysical survey in the northern shelf was carried out to collect single-channel shallow seismic data, gravity cores and surface sediment sampling for studying Quaternary geology in the mid-1990s (Oktay et al. 2002). The second stage of data collection was after the disastrous earthquakes of 1999. A strong international effort that followed the 1999 Izmit earthquake that culminated with the MarNaut mission, where several dives were devoted to the study of gravitational deposits (Henry et al., 2008; Zitter et al., submitted). Then, although specific studies on fossil landslides have been carried out in the past (Gorur and Cagatay, 2010; Ozeren et al., 2010) a detailed study of submerged areas, that will be most likely sites of major landslides in the future, is missing.

Another offshore data set collected in 2007 by TÜBİTAK MAM, will be also used to focus on the fragment of the Western Black Sea fault (WBS). It includes high resolution bathymetry (<20m), bathymetry (20 m<H<100 m) and very dense shallow seismic lines along the shoreline (Ergintav et al., 2011). The multichannel seismic data acquisition is carried out for the first time at west of the Bosphorus in the northern shelf of the Sea of Marmara to investigate offshore structural features such as Çatalca Fault Zone and the shallow deformations as the result of possible submarine landslides to interpret the structural features offshore the Avcılar Peninsula (Ergintav et al., 2011).

By the way, ISMAR-CNR collected, during several expeditions starting from 2001, multi-beam bathymetry, high-resolution single- and multichannel seismic reflection data, as well as gravity cores on a major potential submarine landslide (4 x 2 x 0.2 km) located at the entrance of the Izmit Gulf, close to the W termination of the surface rupture of the 1999 Izmit earthquake (Gasperini et al., 2011). The dimensions and characteristics of the landslide, together with its location (close to the NAF northern strand, and facing the Istanbul coastline) are important reasons for attempting a detailed study of this major landslide, through already available geophysical data, and carrying out modelling along with ITU to predict its behaviour during the next major earthquakes that will probably affect this strand of the NAF.

Task 1.b On-shore landslides

The area of Cekmece, consisting essentially of gently rolling hills with low slopes, is covered by more than 400 landslides showing numerous scarps (Duman et al., 2005). Approximately half of all landslides are distributed between Büyükçekmece and Gürpınar area of the Avcılar Peninsula, which are important local landforms in the region.

Field investigations and analysis of this area have been published, delivering existing inventories, available GIS and existing monitoring of local block deformations with Global Positioning System (GPS). A deeper analysis will be undertaken (distribution, density and activity analysis of landslides, spatial persistence and temporal frequency of landslides) with further field survey for investigating landslides and evaluating the evolution of existing mass movements of much concern. Development of landslide hazard map and associated uncertainties will be carried out including strong ground motion along with estimated amplification site effects, dealing also with intense and/or prolonged precipitation as a potential worsening factor during seismic shaking. While carrying out geophysical investigations in relation with Task 2, an active landslide will be identified as a potential pilot site to be instrumented in the future with a specifically designed ground-based system for local continuous and multi-frequency observation to study physical interactions and early warning strategies.

Use of space multispectral/hyperspectral image data to identify geological and geophysical parameters and delineate corresponding areas will be completed to evaluate the resolution to identify landslide hazard-related features. Evaluation and fusion of the extracted features with InSAR related ones and geological/geophysical models should permit to design a suitable strategy to help defining a landslide hazard scale.

Moreover, the integration of geological and geomorphological analyses with high-resolution DInSAR data will allow the identification and characterization of activated and reactivated Deep-seated Gravitational Slope Deformations (DGSD). The modelling of the related deformations will permit to characterize from a geometrical point of view the sliding plane and to quantify the amount of slip.

Eventually, guidelines for an aggregation strategy between field surveys, ground-based and space geological and geophysical data will be produced to refine a regional landslide hazard scale to be used.

Task 2. Ground motion data, local seismic site effects and dynamic numerical modelling 

Task 2.a Off-shore landslide and Tsunami hazard

The numerical modelling and laboratory testing of landslide generated Tsunami scenarios in the Sea of Marmara will be an ITU contribution. Collaborations between ISMAR and ITU are already under way to study a sediment mass at the entrance of the Izmit Bay. Several seismic images of the mass are being studied (Postacioglu and Özeren, 2008). A 3D generation model will be assimilated into a shallow-water finite-element Tsunami propagation code being developed at ITU. However, ITU will carry out numerical simulations of tsunamis generated by a possible mobilization of this mass. Furthermore, run-up scenarios will be produced by using the outputs from the numerical models. A 15 m long and 60 cm wide and 1.5 m high tsunami channel has been constructed at ITU hydraulics lab to this end. Especially 1-D run-up scenarios will be tested in this channel (Özeren and Postacioglu, 2012).

Task 2.b On-shore landslides

The Ataköy area, also located in the same western part of Istanbul, has strongly been affected by the Mw=7.4 Izmit Earthquake although it is about 100 km away. Local site effects played obviously an important role to increase the damage together with the bad building stock. This has been confirmed from field studies and seismic data collected sometimes after the disaster.

The assessment of ground motion reference scenarios and local seismic amplification effect is possible by in-situ data acquisition and processing. Site effect studies (e.g. mapping of predominant frequencies and bedrock depth distribution, site amplification), 1D and 2D Vs structure around the north part of the NAFZ, specifically in the western part of Istanbul and its suburbs, will be studied. This will include compilation of strong motion data from the large earthquakes (Mw>5) in and around Marmara Sea, and the recent studies on strong ground motion and site effects, which have been performed in the area.

A series of geophysical surveys (determination of S-wave velocity structure at depth with active/passive array surface wave measurements) and geological investigations carried out by TÜBİTAK MAM, to obtain regional information on the macro scale bedrock properties with depth (Ergintav et al., 2011) will be      considered to plan a new and pointed local geophysical study of the area.

A new campaign of ambient microtremor recording to assess in a cheap and fast way the fundamental resonance frequency of a given site, based on Horizontal-to-Vertical Spectral Ratio (HVSR) from single-station measurements (also known as ”Nakamura method”) (Nakamura 1989, Nakamura 2000, Lermo & Chávez-García 1993). The resulting spectral ratio gives frequency dependent amplification for the site.

Previous study on local site effects give evidence that the Avcilar district of western Istanbul (Özel et al. 2002, Tezcan et al. 2002) is characterized by the presence of soft sediments in basin structures and this has caused strong amplification of earthquake ground motion during past earthquakes. The alluvium, on the other hand, represents the most critical unit in terms of site amplifications and is limited to the fluvial depositional centres. The gentle topography of the area, with shallow synclines and anticlines plunging towards the Marmara Sea in the south, represents an environment significantly different from classical alluvial valleys or closed sedimentary basins. In this respect, the expected site effects also differ significantly (Sørensen et al., 2006).

The microtremor campaign will be carried out also to identify the best locations to set up a temporary local seismic network to integrate the Marmara Seismology Network of Turdep.

A local seismological network is needed to fill the gaps of available seismological catalogues to catch any microseismological activity (Ergintav et al., 2011). Then, a complementary local seismic network, composed of a few broadband accelerographs, will be temporary installed to enable a good focus on this area. Following the Seismicity Map and Earthquake Density Map of the USGS Earthquake Hazard Program for the Ataköy area it will be highly probably to record earthquakes (from small to moderate) in the near and far field conditions during two years of seismic monitoring. Two stations could be installed inside and outside the landslide area chosen as pilot site, a station on the limestone outcrop (if possible) as reference station and the other two stations on alluvial deposits with known lateral variations.

Thus, the site effects will be expressed in terms of amplification describing the ratio of the ground motion at the free surface to that at bedrock level. For local variations of site effect H/V spectral ratios are calculated for recorded microtremor data.

All geological-geotechnical derived from task 1 and seismic data available (from permanent existing stations and complementary local seismic array) concerning the selected area will aim at pointing out local seismic amplification effects due to geology, topography as well as directivity and polarization of seismic waves. A preliminary geological model will be defined on the basis of the available data to conduct 1D and 2D linear numerical modelling aiming at analysing the role of topographic and stratigraphic conditions (including effect of incidence, directivity and polarization) on the surface shaking. The local in-situ measurements and the numerical results will be extrapolated to give a good estimate of the amplification in the areas where only sparse data are available nowadays.

Engineering-geological models will be mainly defined based on the already available data to depict the landslide geometries and to define the more adapt rheologies to be considered for the involved soils. This will allow performing dynamic numerical simulations in nonlinear conditions devoted to: i) back-analyze historical events and sequences of monitoring records for validating and calibrating the engineering-geological models, ii) evaluate the role of geological features on permanent deformations as well as on landslide triggering; iii) quantify the effect of different seismic inputs. Most results of Task 2 will provide input into Task 1.