Paleotsunami record of the past 4300 years in the complex coastal lake system of Lake Cucao, Chiloé Island, south central Chile

Abstract

In 1960 CE, Lake Cucao on Chiloé Island in south central Chile was inundated by the tsunami of the Great Chilean Earthquake (M_w 9.5). The area of what is now the lake basin has been submerged since the end of the rapid postglacial sea-level rise and has recorded tsunami inundations in its sediment record since then. This study reconstructs the tsunami history of Lake Cucao. Reflection-seismic profiles and side scan sonar data of the lake reveal a tidal delta with a crosscutting channel, which controls the sedimentary environment in the coast-facing part of Lake Cucao. The convergent pattern of seismic reflections near this channel indicates that tidal currents were active in the lake at least episodically since the formation of a major unconformity with strong reflection amplitude, which records the onset of lacustrine sedimentation. A radiocarbon date at the base of one of the 21 collected sediment cores dates this reflector to ~3800 years BP. Little net vertical displacement (≤2 m) in combination with an outlet river channel that can act as a pathway for sediment transport appears to have maintained the sensitivity of Lake Cucao to record tsunami inundation. The sedimentary record contains 15 clastic layers which are interpreted as tsunami deposits. The confidence level on the tsunami interpretation depends on five site-specific criteria, which are: (i) high magnetic susceptibility of the sediment indicating high clastic content, (ii) cross core correlation indicating widespread deposition, (iii) acoustic reflector correlation to the sedimentary record (also indicating widespread deposition), (iv) presence of mud clasts, and (v) age correlation to known paleotsunamis in the area. In this way, eight clastic layers are interpreted as tsunami deposits with a high confidence level, five with a medium confidence level and two with a relatively low confidence level. This study adds a long paleotsunami record on a coastline where extreme tsunamis occur frequently and where long (>2000 years) paleotsunami records are still sparse.

Introduction

Long and continuous sedimentary records of infrequent large-scale tsunamis are essential in characterising recurrence patterns – a prerequisite for reliable hazard assessments. In recent decades the scientific means to study the sedimentary record of tsunamis has grown in quantity and quality (e.g., Chagué-Goff et al., 2011, Chagué-Goff et al., 2017). Linking tsunami deposits to documented historical tsunamis is a necessary step to calibrate tools in paleotsunami research (e.g., Costa et al., 2016). However, the primary reason for sedimentological investigations is to extend the historical record which is often not long enough to capture the variability in tsunami recurrence (Kempf et al., 2018). This is a challenging task because long and continuous sedimentary records in often highly dynamic coastal areas are rare.

Written history in south central Chile begins with the Spanish invasion in 1535 CE (Lomnitz, 1970, Lomnitz, 2004; Cisternas et al., 2005). In the ~500 years since then, four major earthquakes were chronicled in the area between the Arauco peninsula (~37°S) and the Chile Triple Junction (~46°S). The latest was the 1960 CE Great Chilean Earthquake (M_w 9.5), notorious for being the strongest earthquake on instrumental record. The older events occurred in 1575, 1737 and 1837 CE. According to damage reports, tsunamis damaged coastal towns in all but the 1737 CE earthquake. Sedimentological investigations produced evidence for tsunami inundation for all four documented tsunamis in many coastal areas of Chile (Cisternas et al., 2005, Cisternas et al., 2017; Reinhardt et al., 2010; Atwater et al., 2013; Ely et al., 2014; Dura et al., 2015; Garrett et al., 2015; Kempf et al., 2015; Nentwig et al., 2015). In addition, six sites, Tirúa (Nentwig et al., 2018), Maullín (Cisternas et al., 2005), Caulle (Atwater et al., 2013), Chucalén (Garrett et al., 2015), Cocotué (Cisternas et al., 2017) and Lake Huelde (Kempf et al., 2017) are known to have recorded tsunami inundation before the earliest written reports of earthquake and tsunami damage in 1562 CE (Lomnitz, 2004) (Fig. 1). Of these six sites, only the Maullín and Lake Huelde records extend the tsunami history past 1000 years BP, highlighting the need for adequate sites that have a long sedimentary tsunami record.

One of the difficulties when investigating tsunami deposits on millennial timescales is relative sea-level change, which plays a key role in tsunami deposition and preservation (Kelsey et al., 2015; Dura et al., 2016; Szczuciński, 2020). Relative sea-level rise creates the needed accommodation for tsunami deposit preservation in coastal lowlands. However, with too much relative sea-level rise or fall, coastal erosion or progradation, the shoreline displacement may shift the area of tsunami deposition with respect to previous tsunami deposits, which makes long and continuous paleotsunami records from coastal lowlands rare. In contrast, coastal lakes can provide excess accommodation and may be in a position with a relatively stable sensitivity to record tsunami inundation since the culmination of the early Holocene sea-level rise.

This study partially builds on a chapter in the doctoral thesis of the first author (Kempf, 2016). It presents a long and continuous sedimentary record from coastal Lake Cucao, with multiple tsunami deposits reaching back to ~4300 years BP based on a dense coverage of acoustic data and 21 sediment cores. The quality of the tsunami record is assessed by evaluating the sedimentary environment, tsunami deposit composition and age correlation with paleotsunamis in the region. The spatial multiproxy approach allows us to assign a confidence level to the interpretation as tsunami deposits, allowing a better evaluation of the temporal correlation of paleotsunamis.