Palaeolimnology of Lake Sapanca and identification of historic earthquake signals, Northern Anatolian Fault Zone (Turkey)

Abstract

Lake Sapanca is located on a strand of the Northern Anatolian Fault Zone (NAFZ, Turkey), where a series of strong earthquakes (Ms >6.0) have occurred over the past hundred years. Identifying prehistoric earthquakes in and around Lake Sapanca is key to a better understanding of plate movements along the NAFZ. This study contributes to the development of palaeolimnological tools to identify past earthquakes in Lake Sapanca. To this end several promising proxies were investigated, specifically lithology, magnetic susceptibility, grain size (thin-section and laser analysis), geochemistry, pollen concentration, diatom assemblages, ¹³⁷Cs and ²¹⁰Pb. Sedimentological indicators provided evidence for reworked, turbidite-like or homogeneous facies (event layers) in several short cores (<45 cm). Other indicators of sediment input and the historical chronicles available for the area suggest that three of these event layers likely originated from the AD 1957, 1967 and 1999 earthquakes. Recent changes in sediment deposition and nutrient levels have also been identified, but are probably not related to earthquakes. This study demonstrates that a combination of indicators can be used to recognize earthquake-related event layers in cores that encompass a longer period of time.

Introduction

Movements of the Anatolian and the Eurasian Plates along the Northern Anatolian Fault Zone (NAFZ) in western Turkey have led to several large earthquakes during the 20th century. Epicentres of large earthquakes have moved westward along the NAFZ towards the megacity of Istanbul (Barka, 1997, Stein et al., 1997). Reconstructing the history of past earthquakes contributes to better predicting future earthquakes in the region. Knowledge of past earthquakes in earthquake-prone areas raises the possibility of protecting humans and infrastructure (Atakan et al., 2002, Sørensen et al., 2006). It was for these reasons that the European Union project on “Large Earthquake Faulting and Implications for the Seismic Hazard Assessment in Europe” (RELIEF) was initiated (Pantosti et al., 2008).

Lake Sapanca is an ideal study site, because the lake is located directly above the NAFZ fault at the centre of the RELIEF project region. Palaeoseismological trenches to the east and west of Lake Sapanca have provided evidence for palaeo-earthquakes (Rathje et al., 2000, Hitchcock et al., 2003, Altunel et al., 2004). However, it is important to complement trench investigations with a palaeolimnological study to provide a more complete picture of plate movements along the NAFZ, because movements along one part of the NAFZ may influence movements elsewhere (Lettis et al., 2002). Lake sediments may also contain an archive of past earthquakes, because they are deposited continuously.

Some researchers have successfully identified earthquakes using sedimentology, geochemistry, pollen or diatoms in lake sediments as proxy indicators (Sims, 1975, Ken-Tor et al., 2001, Arnaud et al., 2002). For example, sedimentology has been used to identify in situ deformations (Enzel et al., 2000, Becker et al., 2002, Migowski et al., 2004) and turbidite or homogenite deposits (Monecke et al., 2004, Bertrand et al., 2008) generated by earthquakes. Similarly, earthquake-induced slumping and deposition of turbidites have been identified in lacustrine sediments of Lake Annecy, France (Beck et al., 1996), Lake Biwa, Japan (Inouchi et al., 1996) and Lake Le Bourget, France (Chapron et al., 1999). Geochemistry may indicate whether sediments are allochthonous or autochthonous and thus provide insights into earthquake-related changes in transport and sedimentation processes. The combination of sedimentological and geochemical indicators allows varying environmental changes to be better defined. For example, total organic carbon provides information about organic productivity within the lake, which mainly depends on primary productivity, and about productivity within the catchment (Wagner et al., 2004). Allochthonous organic carbon in a lake varies with changing vegetation cover in the catchment and changing precipitation patterns, which in turn may change vegetation cover and also sediment transport energy (Håkanson and Jansson, 1983, Wagner et al., 2004). In addition, the amount of total inorganic carbon, and changes in element composition, grain size, and magnetic susceptibility can be used as indicators of past environmental changes and earthquakes (Bertrand et al., 2005, Franz et al., 2006). Each lake reacts differently to strong earthquakes, due to its unique bathymetry and catchment characteristics (Becker et al., 2002). Specific earthquake signals first should be identified on short lake sediment cores that span a period with known earthquakes. This information can then be used to infer the earthquake history from long cores spanning thousands of years (e.g. Leroy et al., accepted for publication).

Pollen have also been shown to record past earthquakes (Cowan and McGlone, 1991). For example, uplift or subsidence (Mathewes and Clague, 1994, Mirecki, 1996), or exposure of new habitat for pioneer plants (Cowan and McGlone, 1991) may be evident in fossil pollen records in tectonically active areas. However, the NAFZ at Lake Sapanca is mostly a right-lateral strike-slip fault (Straub et al., 1997), and no extensive areas of uplift, subsidence or exposure of the fault plane would be expected in an earthquake. Therefore, in Lake Sapanca palynological data may provide indirect evidence of earthquakes, largely as a proxy for changes in sediment sources and sedimentation rates. For example, pollen concentrations may indicate whether sediment originated from rapid in-wash of soil, which are usually aerobic and pollen-barren, or by direct accumulation of pollen in deep water sediments, which are often anaerobic and pollen-rich.

Diatoms have been used successfully as palaeo-indicators of earthquakes in coastal waters (Hayward et al., 2004) and are potentially useful in lakes. Seepage of hydrothermal fluids along a fault located below the sea may lead to salinity or pH changes in the water during or after an earthquake (Claesson et al., 2004, Kuşçu et al., 2005). However, the local hydrogeological setting determines whether any change occurs at all (Sneed et al., 2003, Woith et al., 2003). For example, the 1999 earthquakes in northwest Turkey (Fig. 1) changed the conductivity of water in a well ∼1400 km away from the epicentre. However, conductivity levels did not change in wells ∼300–1200 km from the epicentre, due to differences in geology (Woith et al., 2003) and it is not known whether the water chemistry of Lake Sapanca changed during these earthquakes. Another possible earthquake-related limnological change in Lake Sapanca is an increase in nutrient levels caused by mobilisation of nutrients from sediments in the lake or its catchment. Diatoms are generally good palaeo-indicators of water chemistry changes (Stoermer and Smol, 1999) and can be used to identify whether the waters of Lake Sapanca changed following earthquakes.

The aim of this study is to test whether recent sediments deposited in Lake Sapanca contain proxy indicators of earthquakes over the past few thousands of years. Lithology, grain size (thin-section and laser analysis), magnetic susceptibility, geochemistry, pollen concentration, and diatom assemblages were studied for their ability to identify palaeo-earthquakes in Lake Sapanca.