Scribd Logo
Upload

Welcome to Scribd!

  • SSB-SSS 3

    Uploaded by

    Valentina Henriquez Avilez
    11 views14 pages

    Original Title

    SSB-SSS_3

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    11 views14 pages

    SSB-SSS 3

    Uploaded by

    Valentina Henriquez Avilez

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 14
    ‘Contents Hats avaiable at ScioneeD rec Sedimentary Geology journal homepage: www.olsevi comilocate/sedgeo ELSEVIER The sedimentary record of the 1960 tsunami in two coastal lakes on Isla de Chiloé, south central Chile Downe P. Kempf**, J. Moernaut ™*, M. Van Daele*, F, Vermassen *, W. Vandoorne *, M. Pino», R. Urrutia “, S. Schmidt *, E. Garrett #, M. De Batist* 2 Renard Cre Marne Cel (RAG, Ghent Une hee Beg * aro de Conca Unt de Cie (ACR) Va Ce oles stint ED Zh eh Send © Ce EULA Unies de Cone, Coen, he + uss ErOc Urey srw Pos Ponce "Gel Sane Bea Roya Beas ett of ara soy Br, gr #5 eel Reach Un Deparment f Csr, Dahan ena, Dara ARTICLE INFO ABSTRACT eee 5 ft 2015 Reed in esd fan August 2015 hep? Aart 2013 ‘hisstdy describes sediments deposited bythe sonal folowing ie 1960 Great Chilean Earthquake (My 83) ‘nto coastal lakes, Lakes Coca ad Held, onthe wes coast ofl ce Chil, south centeal Chile (42.6 bottom profes and se scan sonar mosaics illustrat the sedimentary context of ransecs of gravity cores The stratigraphy ofboth aks features gta semenacion, interrupted bythe abrupt emplacement ofa sandy ayer ‘with mud rp-sp clasts and a mud cap. This sandy aver elects a sudden change in sedimentary envionen, ‘most probably cause by high-energy inundation, Radionuclide analyses ("sand "PH date the inundation Aeposit shor before the mid 1950s The only kngwn event that matches the sedimentologial an cono= Toglal enters the AD 1980 tsunami Using ran size analysis and comparisons with samples fom ager en- Vionments, we demonstrate tat the proximal (eawaré) par ofthe deposit consists of a mixture of sand erved fom subaera sources and eworked gta lake sediment. nthe dsl landward) part of Lake Cua, ‘he sand component isost and the depost consstsenily of remebilsed ake seimeonts The repetition oft ‘nami depost sequences in Lake Huele suggests ainimum of three inundating waves Sub-botom profiles and ‘Se scan sonar mosaics revel tsunami inundation over the bari and more prominent through the outet Fiver channel. The dominant roe ofthe ver channel sa pathway for sediment transport is also éescribed in cae samples by sunami depos that ine aay trom the chanel mouth. The denifaton and description of ‘the deposi left by a known tsunami provie important insights int tsunars sedimentation in coastal lakes and have the potential to ep inthe sear for paleosunami evidence (©2015 Elsevier BY. Allright reserved aor: ht iano 1950 rea Chilean Earthquake Cains ans 1 Introduction lnvestigations into the sedimentary records of tsunamis predomi nantly focus on subaeral coastal settings, including lowlands, marshes, and swales (eg, Cisternas etal, 2005; Brill etal, 2011; Bahiburg and Spiske, 2012; Szczucifik’ et al, 2012a, 2012; Goto etal, 2014) Partic- ularly along generally subsiding coastlines, these locations have the potential to preserve evidence of multiple tsunamis over hundreds fr thousands of years (eg, Atwater, 1992; Cisternas et al, 2005; [Nanayama etal, 2007: Spske et al, 2013), (MG), ent ety, 000 hes Blam. El res: pol xemprCugentbe (Peg) pedo 10 1016) sdgr.2015 08004 {07-0738 2015 Eheier BV Argh eed sunami deposits on coastal lowlands have been the subject of ex= tensive studies and many oftheir characteristics are well described. They typically consist of coarse-grained sheets extending hundreds of metres to 2 few kilometres inland (eg., Dawson et al, 1995; Goto et al, 2014) with similar quasi-simultaneous deposition along vast areas of coast potentaly exceeding 100 km (eg. of tal, 2010). De- posits may reach afew tens ofcentimetresin thickness ad often thin in 8 landward direction (eg, Goto etal, 2014). Proximal onshore tsunami deposits may consist of grain sizes from mud to boulders (eg, Dawson tal, 1988; Paris etal, 2009; Chagué-Goff etal, 20123; Szcrucifiski eal, 20122), depending on availabe source material, and may include rip-up class reflecting erosion of underlying sediments (eg, Goff etal, 2009; Atwater eta, 2013). Tsunami deposits often have an erosional lower contact (eg. Atwater, 1987; Minoura et al, 1997). Sand and mud particles are principally transporcd in suspension, Their grav- itational settling often results in single or multiple upward fining sequences (Gelfenbaum and Jae, 2003), Tsunami backwash is often ‘more channelised than tsunami run-up; seaward-extending scours commonly form behind boulders in the thalweg; relatively ite ersion and reworking ofthe tsunami run-up deposit may occur outside the thalweg during backwash (Yamada eta, 2014), In contrast to tsunami deposits on coastal lowlands, tsunami de- posits in coastal lakes ate studied less often, although these settings have two principal advantages: 1) coastal lakes often have excess ac- commodation space and a higher preservation potential; and 2) contin- uous sedimentation with relatively stable accumulation rates between tsunami inundations may enable the development of high-resolution age models. However, suitable coastal lakes are rare as the width, height, and geomorphology ofthe barrier separating the lake from the fcean must be such that only infrequent extreme wave events reach the lake, For example, during relative sea level fll the barrier wil be- come ider and the lke level will become higher, possibly resulingina coastal lake, whichisno longer reached by lage wave events, Neverthe- less, coastal lakes and lagoons have provided insights into the sedimen- tary processes assocated with tsunami occurrence in several regions Paar fs) zs! perc ry 1A ry a er enevos [EJauies (revert) Oo rriver bod FaDhigh terrain P Kemal Siena Gg 24 1997-88 eg, Norway (Bondevik etal, 1997), Sri Lanka (Jackson etal, 2014), Japan (Minoura et al, 1994; Sawai et al, 2008), and Cascadia (Hutchinson et a, 1997; Kelsey et al, 2005; Witter etal. 2012) nami deposits in coastal lakes reflect turbulent, high-energy transport and deposition in an atherwise low-energy steady state sed- mentary environment (eg. Kelsey et al, 2005; Sugawara et al, 2008) They are composed of surrounding shallow marine, terrestrial and la- cusiine sediments, eg, dune sand and organic-tich lacustrine mud, De- posits may contain mud rip-up clasts comprising eroded lake Sediments, including pea, gytja, and silt (eg, Bondevik etal, 1997) Resuspended lake sediments are trapped within the lake basin and set- tle slowly to form a mud cap after the coarser fraction is deposited (Minoura eta, 1994). ‘This paperisthe fist to describe the deposit left by the AD 1960 tsu- ‘nami in coastal lakes in Chile, Using a combination of side scan sonar, sub-bottom reflection seismic profiles, and sediment cores, we recon strUct the inundation of two lakes near the village of Cucao on Isla de Chil, south central Chile (Fig. 1) and characterise the resulting tsuna- mi deposit. This study aims to 1) describe the setting ofthe lakes and rey Biases % rent anahouees sch Bitomporate rain oe Fig.) OvriwafSoth Amer) the opographic and bits mape ash central hearse on FTORD état and akin 208) withthe 1940 repre rane ae Nena eta (208))geararpholpal use sedan Sel aberatons and extend sing site apt (Coople Fah) The ke aye mp re bed oe Soa Pe Kee a Siena Gr 242015) 73-86 6 put sediment cores into their sedimentary context using geophysical tools; 2) identity the sedimentary characteristics of the tsunami deposit; ane 3) assess the contributions of subaeral and lacustrine sediment ‘sources tothe tsunami deposit. 2. The AD 1960 earthquake and tsunami (On the 22nd of May 1960, the entire 1000 km long Valdivia segment of the Chilean subduction zone ruptured in the main event of the (My 95) 1960 Great Chilean Earthquake (Cifuentes, 1989). This megathrust earthquake, with sts rupture zone mostly offshore (Barrientos and Ward, 1990; Moreno etal, 2009) (Fig. 1), generated a tsunami with a maximum reported near-field wave height between 15 and 20 m (Sievers etal, 1963), Damage to coastal infrastructure ane oss of ie occured not ony in Chile (Sievers eta, 1953) but also in far-eld regions across the Pacific, including Hawaii (Lander and Lockridge, 1989), New Zealand (Chagué-Coff et al, 2000), French Polynesia (Vitousek, 1963) and Japan (Nanayama et a, 2007). Isia de Clog les in the central parto the 1960 rupture one (Fig. 1) ‘The two closest eyewitness reports ofthe tsunami ate fom lighthouses ‘The chief ofthe Punta Corona lighthouse, about 90 km north of Cucao,, reported 2 15-20 m high wave. Officers ofthe Punta Weather light= house on Isla Guafo, bout 120 km south of Cueao, measured run-up of 10 m (Sievers etal, 1953). No eyewitness reports were obtained from close to our study locations (Fig 1)-Nevertheles, 2 ater sociolog- ical study about life i the village of Cucao describes the 1960 tsunami as hhaving damaged 12 and completely destroyed 27 ofthe 52 buildings that formed the village at the time, including the church, which was located only tens of metres away from the shoreline of Lake Cucao (Weisner, 2003) (Fig, 1). An extensive survey 8 years after the earth- {quake reported that co-seismic subsidence in Cucao was about 1 m (Phtker and Savage, 1970), At Ieast 13 lakes in south central Chile recorded the intense shaking that accompanied the AD 1960 earthquake through a variety of seismi- cally triggeted event deposits (Van Daele et al, 2015). Deposits of, the tsunami and evidence for co-seismic subsidence are described in coastal lowlands located to the north of our study area, ie. at Baha {Quetalmaliue (41.85°S) on Isla de Chiloé (Gartet et al. 2015), inthe Rio Maullin estuary (41.60°S) (Cisternas etal, 2005; Atwater etal, 2013), in Rio Cruces (39.77°S) (Reinhardt et al, 2010), and at Tir (38.34°S) (Garrett etal, 2013; Fly et al, 2014; Nentwig etal, 2015) (he 1). 3. Geomorphic setting ‘This study focuses on two lakes on the west coast af sacle Chilo Lake Cucao (07409°W, 4263'S) and Lake Huelde (074.11°W, 4260'S) (Fig. 1), Lake Cucao is Separated from the ocean by a 1.3 km wide and 5-6 m high barrier; the maximum lake water dept is 21 m, The lake surface is at sea level anda river channel through the barter allows ex- change of water with the ocean, The marine water inflow during high tide is enough to cause a stable saline boctorn-water body throughout the year, despite strong winds and shallow water depth (Vilalobos tal, 2003), A bartier 1.1 km wide and 5-6 m high separates Lake Huelde from the ocean. The lake surface les at an altitude of -1 m and the maximum water depth is 11m, A small steam discharges water tothe ocean; the lake i situated above the intertidal zone and there is no interchange of saline water at present. “The bedrock inthe area consists of metamorphic rocks of Paleozoic and Triassic age, mostly metapelites, metacherts, and metabasites. This ‘metamorphic basement is covered by sandur sediments from the last slacial period (35-141 ka) and its predecessors (SERNAGEOMIN, 2003; Glasser etal, 2008). The geomorphological and landscape units between the ocean and the lakes are beaches, recent and ancient dunes, bogs, riverbeds, freshwater marshes, pastures (protected from wind erosion by man-made 3-5 m high sand ridges), dense shrubs, and temperate rainforest (Fig. 1). Forest covers more than 70% of the Catchments of oth lakes. The catchment-area-to-ake-area ratios of Lake Cucao and Lake Huelde are 3.1 (Villalobos etal, 2003) and ~75, respectively, 4 Methods Ad. Acoustic imaging In order to get detailed spatial information about the lake floor ‘morphology and acoustic reflectivity (a function of sediment hardness and roughness), we collected side scan sonar data in Lake Huelde and in the western part of Lake Cuczo. The data were obtained with a kein3000 side scan sonar, which utilises frequencies of 100 andl 500 kz and produces a 50 m wide swath. The resulting horizontal resolution is 8-40 cm depending on the beam angle, Positioning was recorded with a Garmin GPS antenna with a 3 my horizontal error. Navigation smoothening, slant range correction, and mosaic selection ‘were performed using SonarWieMap v.4. We also excraced the bottom track ofthe sie scan sonar data to create bathymetric maps for the lakes using a sound velocity of 1480 m s~! High-resolution seismic reflection profiles were acquired in the western patt of Lake Cucao and the whole of Lake Huelde, using a Geopulse pinger (Frequency ~3.5 ki). Lake Cucao and Lake Huelde were surveyed with 71 and 31 lines with a total of25 km and 23 km with a line spacing of 150-200 m and 200 m, respectively, in areas of interest. These seismic data image the stratigraphy ofthe lacustrine sediment infill ata vertical resolution between 10 and 20 cm, Profiles ‘were processed (bandpass filtering) and interpreted using IHS Kingdom, Suite v.88, 42, Suberil and lacustrine sediment samples 0 identify tsunami deposits and thelr source se¢iments and t0 distinguish tsunami deposits from regular lacustrine sediments, we sampled both the barrier between the lakes and the ocean and the sediments within the lakes and analysed their sedimentologicalcharac- teristics, specified below. In 2013, we used 10 cm long plastic tubes (3.0 em diameter) to cok. lect 27 surface samples from the barter area of Lake Cucao (C14-C28) and Lake Huelde (H15-H25) (Fig 1). These samples give an overview ofthe subserial sediments that would be available for transport during extreme wave events at present, Surface samples were taken from the beach, cune, freshwater marsh, and other subaeral environments that, lie between the lakes ané the acean, Gravity cores were collected from Lake Cucao (cores C1-C13) and Lake Huelde (cores H11-1114) in 2011 and 2012 (Fig. 2). These coreshave alength of up to 1 mand an inner diameter of 6.0 crn. Cor- ing locations were selected based on acoustic imagery data, 43, Sediment analysis All subaerial samples were analysed for magnetic susceptibility using aBartington MS2E point sensor. The mineralogy was determined under a stereoscopic microscope. Because the higher magnetic suscep- tibily values correlate with coarser and heavier minerals (ion oxides density ~5.2 g em™*), we subsampled lake gravity cores 3-5 times with 2 05 cm depth interval around the highest magnetic susceptibility values for grain size analysis. The grain sizeof subaerial and lacustrine Samples was measured with laser granulometry using a Malvern Mastesszer 2000 ate removing the organic content with hydrogen pet- ‘oxide, The maximum grain size that can be measured by the Mastersizer 2000 is 2 ram ‘The lacustrine gravity cores were analysed with a Geotek multi- sensor core logger (MSCL) to measure gamma-ray attenuation density and magnetic susceptibility (same point sensor as above) at a down a) high rotectviy side seams \ outline ‘ial deka {oct af slope —~ rosea basin tore bation « love-snapes higeretecity "Lake Cucao 2am or) aide eda Cg anti open mn) a) epee he hl ina ad tram sin deposi a ake Hose core resolution of 2 mm. A line-scan camera mounted on the MSCI. provided high-resolution spit-core surface images. A Siemens Flash ‘medical x-ray computed tomography (CT) scanner was used to provide 30 views for five cores (voxel sizeof 0.15 x 0.15 x 0 mm, with the 06 mm in down-core direction), Additional classic x-radiography visualised the sedimentary structures in core H11, which was also selected for dating Organic matter was removed and grain size distribu- tions measured with the same methods as for the subaerial surface samples. To measure water and organic matter content ofthe lake sed- iments, twelve samples fom ake gravity cores at depths between 1 and 32 emand a weight of 2-4 g of wet sediment were analysed using loss ‘on ignition after Heir etal. (2001), For age control, "Cs", and 2Ra-activities were analysed incore CB and HII with a low-background, high-efficiency well-type germanium detector (Ge volume of 260 cm’, low-noise Aland Cu com 5) equipped with a Cryo-Cycle (CANBERRA) (Schmidt etal, atthe University of Bordeaux. In core C8 radionuclide activity was measured on 8 samples in core H11 on 17 samples. Activities are expressed in mBq g~! and errs are based on 1 standard deviation. Ex- cess lead (""Pb,.) was calculated by subtracting the activity supported Pe Kee a Siena Gr 242015) 73-86 ” byits parent isotope, Ra, fom the ttal2"°Pb activity in the sediment. ‘We used the constant-flux, constant-sedimentation (CFCS) model (Robbins, 1978) excluding the depth interval of the tsunami deposit to derive a?"°Pb-based age-depth model 5. Results and interpretation 5.1. Side scan sonar data “The side scan sonar mosaics from Lake Huelde and from the western part of Lake Cucao (Fi. 2) show an acoustic facies of high reflectivity on lake basin slopes and on ~2 shallow plateaus adjacent to the lake out- flows, suggesting coarse sediments ac the surface in these areas (bright ‘yellow colour, Fig. 2a, c). Sediments in the central basins below ~13 m. water depth in Lake Cucao and ~7 in Lake Huelde present a faces of loge reflectivity (dark colour, Fig. 2a, ). Low reflectivity corresponds to smooth, fine, and soft sediment surfaces. In both lakes, the transition {om high reflectivity on the slopes to low reflectivity in the basins is sharp. The line dividing the two areas generally coincides with the feat of the slope. 5111. ide scan sonar map of Lake Cucao Inthe west of ake Cucao athe outlet, a channel of upto 100 mwide and 5 m deep dissects a 08 km wide platform, The plateau bears ‘megatipples onan otherwise smooth surface, there are 217 megaripples with an average nipple length of 8 m from crest to crest, approximately 1.5.m high and oriented concentrically around the channel (Fi 2, Their orientation suggests thatthe megaripples ae probably formed by ‘ood currents in shallow water (ef, McCave and Geiser, 1978).As the ‘egatippes still appear fresh and have not been eroded (eg. by wind- ineiuced waves), the plateau can be regarded as an active tidal delta, In the southeast of the surveyed area, a lobe-shaped area at the channel mouth shows high reflectivity (Fig. 2) and isa bathymetric 4epression, Due to its poston and its erosional character, the lobe-shaped area i mast likely an extension of the channel system into the basin. ‘Around the inferred continuation of the channel, there are nu- merous 1-6 m wide high-reflectivity features (Fig. 2c. d). The istr- bution of these superficial or near-superficil features i limited to the west ofthe basin, During the coring carmpalgn, anchors caught grasses and bushes, similar to the contemporary vegetation on the bartier. These accumulations of vegetation possibly caused the bigh-refletivity features 5.12. Side scan sonar map of Lake Huelde ‘A200 m wide plateau with ~2 m water depth atthe outlet of lake Huelée bears no geomorphological features except for a ~50 m wide crosscuting channel in che south linked tothe outlet river (Fg. 23). “Two groups of sub-paallel curved linear features lea the northeastern ledge ofthe plateau. The linear Features are 30-50 m long and 1-3 m ide. The interpretation of these features as rotational landslide scarps {s supported by intact rotated sediment blocks on sub-bottom profiles and i discussed further in Section 52. ‘Anarea ofhigh reflectivity is located atthe mouth ofthe crosscutting channel atthe plateau, Itis-250 m long and 150 m wide with the ong axis oriented from west to east. The positive morphology of this feature stiggests that, unlike the lobate shape in Lake Cuca, thisisadeposition- al feature, ie, a fan. However, the strong signal wipe-out at higher Swath angles suggests thatthe sedimentary body, whichis praducing the high-reflectvity inthe lower swath angles, s buried underneath 3 sot, acoustically transparent sediment cover. This means thatthe side Scan sonar may have directly imaged an acoustically harder layer buried beneath the present sediment surface. Numerous 1-7 m wide high- reflectivity features occur within the otherwise ow-refectivity asin. Close tothe western shore and north ofthe plateau these bright spots ae less jagged and more rounded (Fig. 2), Flevated reflectivity values are observed in the northeast of Lake Huelde, This area is adjacent tothe only noteworthy fluvial input into the lake. Stronger reflectivity is most likely caused by this errigenous input. 5.2. Sub-bottom profiles The sub-bottom seismic profiles reveal the sedimentary architecture of Lake Cucao and Lake Huele (Fig 3).On shallow plateaus in the west ‘of bo lakes, penetration of the sub-bottom profile signal was strongly reduce, probably caused by the presence of coarse-grained sediment. In the deeper basins, the acoustic signal is absorbed after short travel inthe soft sediment, due ro the presence of shallow gas likely a product of organic matter decomposition), While there i ~200 m wide acous- tie window into basin sediments along the gentle western slope of Lake CCucaa, in Lake Huelde penetration below the first strong reflector is scarce However, we define three seismic units common to both lakes, ‘which we refer to as units Ula, Utb, and U2. Uta’s upper boundary isa sharp, hummocky unconformity in the basin. The lower boundary isnot defined. Internally, Ua is made up of strong, wavy, and continuous reflections, with strongly decreasing amnplitude and continuity with depth (Fig. 2a, b). Ua is interpreted as Sandu sediments, which buld the base of both lakes. Unb s similar to Uta in most aspects, including the abundance of strong continuous reflections an the lack ofa defined lower boundary. Utb occurs in shallow water. Unlike Uta, Utb has internal unconfor- mites and its upper boundary is the lake for (Fig, 3) Seismic penetra- tion is rarely more than 3 m sub-surface, Where penetration is higher, strong hummocky and discontinuous reflections are the lowermast vs~ ible feature, Acoustic ringing effects and the low signal-to-noise ratio prevent characterisation ofthe deeper pars. In both lakes, Ub includes eroded channels that cross cut the shallow plateaus inthe west. Tere ate areas of wavy lake floor inthe vicinity of the modem channel in Lake Cucao, caused by the megarippes that are described on the side Scan sonar imagery (Fig. 3a). UIb is interpreted as reworked sandur sediments and other sediments from te barrier, which build the base inthe shallow ateas of bot lakes. (U2 overlies and onlaps Uta wherever the lower contact of U2 is vis ible. Tae most prominent internal signature in U2 consists ofa series of stcong. infrequent, and sub-parallel reflections in both lakes. Between these strong continuous reflections, U2 consists of very low acoustic amplitudes. The top of U2 isthe lake floor, which in Lake Cucao is a strong reflection (Fig. 3a, b), and in Lake Huelde isan exceptionally ‘weak reflection (Fig. 3e), The uppermost strong reflection in Lake Hoelde usually lies between 0.3 and 0.4 m below the lake floor. The same reflector may be responsible for the high reflectivity on low swath angles on side scan sonar data, The reflection is hummocky in the west and transitions gradually toa fa reflection in the east. Within the area of lat reflection the acoustic amplitude decreases eastwards U2 is interpreted as the sediment infil of both lake basins. This infill mostly consists of almost acoustically transparent lake sediments and of sheets of sediment with high acoustic impedance The two groups ofsub-parallel curved linear features identified in the Lake Huelde side sean sonar maps canbe linked with intact rtated sediment blocks and confined wedge-shaped sedimentary bodies seen inthe sub-bottom profiles (Figs 2,3) All ofthese sedimentary bodies are in direct contact withthe uppermost high amplitude reflection af U2. The scarps and rotated blocks in Lake Huclde are therefore ierpreted as simultaneously sriggered rotational lansides with com- pressional toes, Such simultaneously triggered slope failures and land- slide deposits in lakes are usually attributed to earthquake triggering and to local earthquake intensities of VI or higher (Monecke et a, 2004; Strasser et al, 2013; Moernaut et al, 2014; Van Daele et al, 2015). The run-out of these landslides is minimal (10-40 m) and their volume is insufficient to cause lake-wide sheet-like deposits chat Would be imaged as high amplitude reflections on sub-bottom profes. ” Perea Sedimentary Clg 282018) 79-86 1g0__290 250 sq a50 490450 sgo__ goa G0 790 Inke floor surface ims) {weak recton) snl Hs canoes, J 1 0 apoyo go ag gogo 60 Seamertreverknginshalew areas Lake Curso rutile Feiment owsrkng [ake Gucao ‘nano Fig. 2 Pinger sb-batiom filer fae Hic in} ana Ces in) and nate aes here arses non-poor reorkng ofan th saw less ‘hess on the upper spe Seis use vie cloud serpent ves (Ul: ule Ul gee U2 re) hat aes location ane pee ae epee 8 thik bscvertca ns In ake Bd, tee Tot (gees be. deed ine nase) na weak ecto incamparsen wih he suri ete 1960 Guna pet (belie) 9 [econ (bane the 196 sam src a0 cls teak lao tbe dings eee ne) 53. Statigraphy and sedimentology ‘The sediment contains up to 80 wt water; ofthe dry material, up to 40 wt is organic. The organic matter consists mostiy 5.3.1, Subuerial sediments of unidentifiable fragments, fbrous plant fragments (eg. ‘All ofthe subserial samples from both barriers Fig. 1) consist of well culms and leaf veins), diatoms, pollen, and seeds. In Lake Cucao, sorted, fine to medium sand, The modal grain size varies from 224 to the ofganic component also includes periastraca of bivalves, 430 um (© = 216-122). The sand is made of ecual amounts of mineral possibly Diplodan chilenss. The homogeneous mud is easily grains and lithic grains. The mineral are orthoclase plagioclase, quart deformed and has an average gamma attenuation density of iron oxides, epidote, hombendle, and rarely zircon. Magnetic suscept- 13 g em™>. The grain size distribution is symmetrical and is bility is generally high at »250 10-2 SI (mean 780 10 sl) with ex- poerly to very poorly sorted. The mode is 15 um (« = 6.05) on treme values of 3000 10-* St average, and particle sizes range from clay to medium sil ‘This lithofaces has a low magnetic susceptibility between 0 532, Locustrne sediments and 40 107 SL ‘The sediments retrieved inthe 27 gravity cores from Lakes Cucao aminoted gytia—Dark brown, organc-rch laminated mud oc- and Huelde can be subeivided into sx main sedimentary facies (ig 4) curs only in Lake Huelde sediments. Where it occurs, it is invar- ably atthe top of the gravity cores. The itegular millimetre |) Homogeneous gyttaBrown to black organic-rich homogeneous to centimetre-scale laminations consist of yellowish grey and ‘mud is the most abundant sediment facies in te lake cores. black layers within the generally Drown sediment. The Kerra Seteay Gg 328 (2015) 73-85 Inotcae Iho description crsean mage gey bows gey shar lowes conse! and harp upper conae, beeen O'S and Sm tick area spit core sutace image ac brown lace ‘ergarievieh; madam sito tin band he ve |ee)ze\1e oe] er ze ue 0 wz 02 roassive sand with a sharp, sien oresional lower coniaet ‘and ether sharp upper conc fo background {edmont or 9 sharp 0 (raneeonalconnetto 8 mud ea sr ve ce ze 1c 0% ‘rk brown organi ich mug chaste isto Sem n aroter na main of and, sometmes aigned in ayer (onlyin axe Huskie) 1 [es za\ 1s 09 818 dark brown, organievich (enlyin tate Hulde) ip minated| am st brown oback rgnie-reh homogeneous mud 9 homogeneous = 2 a i 4. esc ofthe sediment fac an as asin th the splice sce mae ans compute tomer (C7 ado other physical properties and the organic content are identical to those described fr the homogeneous gyth }) Mu rip-up clasts—Dark brown organic-rich mud clasts of sizes ranging from 0.5 em to larger than the inner core diameter (60.em) occur within a matrix of sand. The mud clasts, which are composed of gytja, ae angularin shape (Fi. 4) and some times aligned in horizons, This facies occurs only in Lake Hivelde and predominanty in cores closer tothe western shore. Becaise ofthe gytja material and the angular shape, they are interpreted as mod rip-up clasts, representing a high-energy fow regime. _Massive sand Massive sands in the lake cores are wellto moder= ately well sorted and can contain finely dispersed organic matter. ‘The grain size distribution consists of a narrow symmetrical peak around 205 yn (¢ = 228); however, 6 out of 11 samples have a very low secondary peak of sit-sized grains. Mineral and thie grains ae equally abundant inthe sand, The mineral gains have the same composition asthe subaeral sands i. orthoclase, plagioclase, quartz ron oxides, epidote, hornblende, and rarely zircon. Strongly fertimagnetic iron oxides produce positive peaks in magnetic susceptibility with a mean of 500 10~* ST nd up to 1750 107? Si, Massive sand units often have a sharp Bnd sometimes erosional contact tothe underlying homogenis ‘vita (facies). In Lake Huelde, massive sand units often occur in association with a mud cap on top (facies vi) and must rip-up Class ({acies i) towards the base. Such sands are transported ina high-energy environment. Y) Organie-rch sit and sand—The dark brown to black, organic-ich, medium sit to fine sand is sometimes indistinguishable from ry Perea Sedimentary Cel 282018) 73-86 homogeneous gytia in colour and consistency; however, ffer- entiation is possible through the higher magnetic susceptibility (60 10% Sh and the coarser grain size distribution, with a mode of 25 wm (p= 531). Out of 19 samples, 9 ate bimodal with a second mode around 182 ym ( = 246). The dominance ofthese two mades varies strongly with location, vi) Mud cop—Mid grey to brownish grey organicrich sit ayers 2 transitional or sharp lower contact and a sharp upper contact. The sil layers aze between 05 and 5:0 cm thick and always occur in association with an underlying massive sand layer. The ‘magnetic susceptibility is low («50 10~° I). The mud cap shares similarities wth the gyttjaof facies and iin organic content and ragneti susceptibility, 54. Facies associations We group the described lihfacis into 2 facies associations (Fig. 4) sya and tsunami deposit. The main criterion is the energy associated with the deposition ofeach facies. Gta including lthotaces i and i reflects the gradual deposition of fine grained maceral in a low-energy eavironment with an organic faction, including plant fragments, dia- ‘toms, pollen and seeds, an litle lati input fom the vvers or wind. The mud rip-up class in sandy matrix (aces i), medium to fine massive sand (Tacies iv), organicrich silt and sand (Facies v), and mad ap (facies vi) form a high flow energy facies association. Apart from the mud cap, the required flaw energy for producing such facies is ‘markealy higher than the low-energy depositional envionment assoc- ated with gytja The erosional contact of massive sand or mud rip-up ‘lasts ina sandy matrix withthe underlying gytja and the mud cap ‘mark the sudden changes from a low- to a high-energy flow regime and back to a low-energy flow regime, respectively. The group of, Sedimentary facies i to vi is interpreted asa tsunami deposit facies association (Figs, 4,5), because this sudden change to extreme flow energy in this coastal lake environment on the Chilean subduction zone in combination with the described sedimentary facies may be linked to tsunami 55. Core chronology In Lake Cucao, the inferred tsunami deposit consistently les only a few ceatimetres below the lake floor. Because the lake floor reflection on the sub-bottom profiles is strong. we link the andy and sity material ineach of the cores with tis song reflector. "Pb, and ”Cs profiles cannot be used to precisely constrain the age ofthe sane layer in core (8. The gytt2 above the sand is only 2 cm thick, leaving room for only two samples (Fig. 6)- However, activity levels of "Cs in the two sam- ples above the sand are higher than in all samples below, indicating epi # @ ‘ in " sof "too ON? ‘e f Ra Se 050 080 om aot mage <0 200 » ow wcemayaaes HT O50 { (Ea wwe FESS] eomicch tandeane Shoei BL pes (R190 tuna cops Fig 5 Tasecsof sr corso ate Coco pper ro) nak ele (ero) Themageisscepity (white ie) platedon tape cae inage—notethe direc 2 Kempo Sete Gg 3282015) 73-85 ® "Pb age-model (year AD) 3920 19602000 ope 10 5 20 as. sop. 35. 404+ 0 16 0 8 1 18a mag. sus. (10° St "Cs (mBq g") ont to 00 "Pb. (mBq 9") 1119190 1 0 80 160 oo 1 2 3 4 mag.sus.(10°S) "Cs (mBq g") Fig. 6 Reon asses on oes Ht (2) and (0). "Cs vals (ack sbow the cs ean It between 1 sn 1 em depth The Pps (rd) the ass forthe ae depth nde (ae) 8B te Cs aves awe deel above the ‘sur depos (re brand neglgble below. ineting the 195s peak to be thatthe sand layer probably predate the mid 1960s. The radionuclide activity in core C8 i too low to allow a robust "Pb age-deplh model Inall Lake Huelde core, a tsunami deposi es at 2141 em depth below the lake floor. This matches well withthe depth of 04-05 ms two-way travel time (30-37.5 em) of the uppermost strong seismic reflection. Therefore, the uppermost strong reflector in Lake Huelde is interpreted as the surface of his sandy deposit Tn core H11, #°Pbg activities present a rapid decrease from 77 mBq g* athe core opto below 10 mBq g~' at about 30cm. asex- pected due to the radioactive decay of Pb (Fig. 6). Considering the sudden deposition at about 20-25 cm and the low activity of "Pb. in depth, we applied a constant-flux-constant-sedimentation model only co the upper part ofthe Pb, profile. The mean sediment accu- mulation rate is 0.39 cm y-, implying an age of AD 1961 = 5 at 20 cm depth, ie, just above the tsunami deposit. "cs profile in cre HI presents a peak ata depth becween 13, and 17 em (12mBqg~"),Jow values near the top (5 mBq ~"), and neg- ligble activities below te tsunami deposit (<2 mBq g~"). The "Cs peak corresponds to the mid 1960s, according tothe annually measured atmospheric fallout of 7Cs recorded at Buenos Aires, Argentina (Quintana, 2011), and sol profiles from the vicinity of Puerto Mont, 20 km north of Isla de Chiloé (Schuller etal, 2002). This is in strong agreement with the modelled age from the “Pb age-depth model ‘Asa result ofthe radiometric dating, corabined with the indication of inundating low deposition, we interpret thatthe deposi i the product, ofthe tsunami that followed the 1960 Great Chilean Earthquake, 56 Spatial trends of te tsunami deposit. In the proximal lake cores, the tsunami depesitis up to 61 cm thick. ‘When, asin five Lake Huelde cores (H1, #4, H6,H7 and HO), the base was not retzieved, then te thickness from the top tothe core base is taken as a minimum thickness. The tsunami deposit thickness decreases with increasing distance from the ocean: however, there ino thinning trend in areas beyond -25 km distance from the ocean (—13 km distance from the outflow: Fig. 72). The lowest thickness in the cores Forthe following measurements of magnetic susceptibility and grain size distributions, we used extreme values to represent the strongest signal for the core location, ie, highest magnetic susceptibility and coarsest gain size distribution, Magnetic susceptibility is primarily controlled by the content of ferrimagnetic and heavy ron oxides (hematite and magnetite have a density of 5.2 g em™*), Secondarily, magnetic susceptibility underies a grain size effect (Thompson and Morton, 1978), Le, larger sand grains produce a higher value Both grain size and density influence the capac- ity of erosion, transport, and deposition of grains; therefore, higher magnetic susceptibility values may reflec stronger flow during tsunami inundation. Organic content, inthe form of larger mud clasts, and finer dispersed organic material in sandy beds attenuates the magnetic sus- ceptbiity values. The maximum magnetic susceptibility ofthe tsunami deposit decreases from an extremely high 1750 10"? Sl 1073 107° SL with inereasing distance from the ocean (Fig. 7b). The latter i stil higher than for gytja of facies i and ii (40 10~* SI). However, the de- creasing trend in magnetic susceptibility i widely disperse: ‘Within the complete grain size dataset, there are three groups of sain size distributions (Fig 8), ic, silt, sand, and bimodal silt and sand, The gran size modes within the tsunami deposit group around a fine (silt) and a coarse (sand) mode. The average fine mode is -25 ym (@= 531) and does not change with distance from the ocean (Fig. 7c 4}. This mode is represented in the unimodal, syrametrcal sit samples, ad in the fine part of the bimodal sit and sand samples. The unimodal silt grain size distributions occur exclusively in organicrieh silt (facies v) and resemble the grain size distribution of gytta (Facies | and i). The average coarse mode is 194 um (y = 2.36) and decreases with increasing distance from acean. The exponential fit curve of this negative correlation has an R-value of © 59 (Fig, Te), The R-value in- creases to 0.74, when the grain size modes are plotted against the dis- tance from outflow (Fig. 74), which is the distance of each sample fom an arbitrary point a the lakes’ alow, The coarse mode is repre- sented in unimodal, symmetrical sand samples an in the coarse part of| the bimodal sit and sand samples. The unimodal sand grain size dsti- butions describe the grain sizeof massive sand (facies iv) and organic- rich sand (facies v) and resemble the grain size disiripution of subaeral samples. The most distal sample with a coarse mode (core 6) is located at -13 km distance from outflow: beyond that distance the tsunami epost consists only of silt (Fg. 72), we Perea Sedimentary Cel 282015) 79-86 ca pias Ge ge" ] fie . 2] . yen abe aden an ®) estar fom cca ae STE 5 won riers of a0] . 2 10004 w. ; . ° °) seco distance from ocean (m) os z jortra lamas woos E wl Tevttdor ye tee tent a) disarce fo ten 1000-3 jsubacral lacusiting, eg Ming tad = OTF = eer $F onsen 3 a distance trom outlow (mn) Fi, 7 Spatial ends of) sam depose) maximum mae sep ‘same modes are pote over distance trom the lke outlow. 6. Discussion ‘The two lakes share many similarities, yet the post-1960 sedi- ‘mentation rate is drastically different. Lake Cucao has a low sedi- mentation rate of O.4 mm yr", while Lake Huelde has a high sedimentation rate of ~59 mm ys~', Due to their different catchiment-area-to-lake-area ratios (Lake Cucao: 3.1; Lake Huelde’ 75), Lake Huelde can be expected to have a ~7 times higher sedi- ‘mentation rate than Lake Cucao (Van Daele et al, 2015). In addition, tidal currents may have inhibited sedimentation or even eroded sgyttja in the western parts of Lake Cucao, especially after the AD sediment source: eres en BS — subaerial samples & wed wn i Elona & =S grain size (ym) Fig. Averaged gain een curve hice ines) ofthe singe spe (hin nc) dcr te sure sede of cal sand (Be) ad te er) a el 2s te thee ferent sans ze composts om he Guna epost (ea ak and Slo) Te eaph shows te anton fhe gran soe esrbatone of feta epost amanda carl cmposton ed) 5p coal er sion (eo). 1960 earthquake, which resulted in -1 m of co-seismic subsidence Inthe study area (Plafker and Savage, 1970). 6.1. Sediment source and distribution inthe lakes “The unimoslal silt and unimodal sand distributions in the tsunami deposit resemble the grain size distributions of gyttia and subaerial sand, respectively, so closely (Fg. 8) thatthe latter two are interpreted tobe reworked and transported by the tsunami Moreover, the mineral gical similarity between subserial and tsunami sand supports this interpretation. Both the gyttja and subaerial sand can be seen as end- ‘member sources to the bimadal sift and sand inthe tsunami deposit. ‘The prominence of the fine ané the coarse mode describes the relative contribution of either source. The area ofthe landward thinning trend and the occurence of sand inthe tsunami deposit overlap (Fig. a,c) they are both limited to ~25 km distance from ocean, i, -13 km distance from outiow. The distal tsunami deposit, ie, beyond ~1.3 km distance from outflow, doesnot follow this landward thinning and fining, because the distal tsunami deposit drapes the basin with the thickest accumulations in the deepest areas (eg, in cores C1, C11, H2, and H13). This matches the observations from Bradley Lake, Cascadia, in whic sandy "dstur- bance event" deposits exhibit landward thinning, while thinner, ‘muddy "disturbance event" deposits drape the lake basin, with thicker deposits in the deeper parts (Kelsey et al 2005). The reason forthe lack of thinning inthe Solely gytsa-derived tsunami deposit may be the slow setting velocity of the grains andthe relatively high-energy, turbulent currents mixing the lakes during the tsunami, When the gains cannot settle fast enough, then thinning cannot develop before the entice lake is evenly mixed. This is especially true for small lakes, eg. Lake Huelde and Bradley Lake, Cascadia This subdivide the lake cores into three zones; 1) a proximal zone, Where the tsunami deposit mostly consists of reworked subaerial sands; 2) an intermediate zone, where the tsunami deposit consists of| 2 mixture of both reworked gyttja and reworked subaerial sands, and 3)adistal zone, where no subaeral sand occurs and the tsunami deposit consists entirely of reworked gytja (Fig. 9). ‘The and deposition limit within the lake may provide a site-specific relative measure of tsunamisize, applicable to paleotstnamy deposits in the same sedimentary record. However, at the moment, this hypothesis lacks the data for testing Studies using the sand deposition limits tore- construct tsunami size onshore are prone to song site

    You might also like