Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench

Quaternary Science Reviews xxx (2014) 1e12

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench Katrin Monecke a, *, Caroline K. Templeton a, Willi Finger b, Brian Houston c, Stefan Luthi d, Brian G. McAdoo e, Ella Meilianda f, Joep E.A. Storms d, Dirk-Jan Walstra d, g, Razali Amna f, Neil Hood c, Francis J. Karmanocky III h, Nurjanah f, Ibnu Rusydy f, Sam Unggul Sudrajat i a

Wellesley College, Wellesley, MA, USA Swiss Agency for Development and Cooperation, Zurich, Switzerland University of Pittsburgh, Johnstown, PA, USA d Delft University of Technology, Delft, The Netherlands e Yale-NUS College, Singapore f Tsunami and Disaster Mitigation Research Center, Banda Aceh, Indonesia g Deltares, Delft, The Netherlands h West Virginia University, Morgantown, WV, USA i United Nations Development Program, Jakarta, Indonesia b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 June 2014 Received in revised form 8 October 2014 Accepted 16 October 2014 Available online xxx

The morphology of beach ridge plains along active margins can be used to reconstruct coastal subsidence during large megathrust earthquakes. Here we use satellite imagery and automatic level surveys to reconstruct the build-up of a new beach ridge along a 10 km long stretch of the western Acehnese coast after the complete destruction of the beach during the great Sumatra -Andaman earthquake and successive tsunami in December 2004. The western Acehnese coast is characterized by ridge and swale topography reflecting the long-term progradation of the coastline. Radiocarbon dates obtained from marshy deposits in between ridges indicate an average progradation rate of 1.3e1.8 m per year over the last 1000 years. As a result of coseismic subsidence of 0.5e1 m and tsunami inundation in 2004, the most seaward beach ridge was destroyed and the coastline receded on average 110 m landward representing 65e85 years of average progradation. However, by 2006 a new 22 m wide ridge had formed. In the following years the coast prograded by an additional 30 m, but has not yet recovered to its pre-December 2004 position. In addition to the spatial data, topographic surveys conducted in 2009, 2012 and 2013 indicate that the crest of the newly formed beach ridge is 0.8e1.3 m higher than the crests of older beach ridges further inland. The source material for the new ridge is most likely sand transported seaward by the back flow of the 2004 tsunami and stored on the upper shoreface. In the months and years after the tsunami, this sediment is reworked by regular coastal processes and transported back to shore, leading to the reconstruction of a higher beach ridge in equilibrium with the vertical displacement of the coast and the resulting higher relative sea level. The preservation potential of the newly formed ridge depends on sediment availability within the coastal system to balance coastal profile adjustments due to rapid postseismic uplift. In Aceh, the preservation of seismically modified beach ridge morphology seems likely and another prominent ridge can be found in 640 m distance to the shoreline. It most likely formed in the aftermath of a previous megathrust earthquake and tsunami about 600 years ago matching sediment and coral records for this region. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Paleoseismology Coastal morphology Beach ridge plain December 2004 Sumatra-Andaman earthquake 2004 Indian Ocean tsunami Aceh (Indonesia)

1. Introduction

* Corresponding author. E-mail address: [email protected] (K. Monecke).

Growing coastal communities around the world are increasingly threatened by extreme events as recently witnessed when Typhoon Hayian hit the Philippines and portions of Southeast Asia in November 2013 or, when the March 2011 Tohoku-oki tsunami

http://dx.doi.org/10.1016/j.quascirev.2014.10.014 0277-3791/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

2

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

struck Eastern Japan and coastal areas throughout the Pacific basin. Using satellite imagery, detailed ground surveying and sediment analysis, the immediate impacts on the coastal zone have been well documented for recent extreme events allowing the quantification of hydrodynamic parameters and geomorphic change with possible implications for storm surge forecasting (e.g. Fritz et al., 2007; Spencer et al., 2014) and tsunami hazard assessment, warning and preparedness (e.g. Jaffe et al., 2006; Vargas et al., 2011; Tappin et al., 2012). Only very few studies have focused on the longer term coastal recovery after extreme events, which is equally important since coastal planners have to allocate resources and design appropriate structures for future hazard mitigation. The partial or complete rebuilding of sandy beaches after large storm events is site specific and has been observed to take in the order of a few months depending on inherited hydrologic and sedimentary parameters (e.g. Wang et al., 2006; Yu et al., 2013). The recovery of sandy shorelines and establishment of equilibrium after devastating tsunamis might take from a few months up to a few years (e.g. Choowong et al., 2009; Liew et al., 2010) and can be expected to be more complex in tectonically active regions experiencing coand postseismic land level changes (Meilianda et al., 2010). In order to prepare coastal communities for future hazards it is critical to understand the frequency and magnitude of past events. Numerous studies of deposits in coastal areas have revealed millennia spanning records of coseismic land level changes (e.g. Atwater, 1987; Shennan et al., 2014) and tsunami inundation (e.g. Nanayama et al., 2003; Cisternas et al., 2005). Latter studies have often been successful in marshy beach ridge plains where tsunami sand sheets have been preserved in low-lying depressions (e.g. Pinegina and Bourgeois, 2001; Jankaew et al., 2008; Atwater et al., 2013). In this paper we combine spatial imagery analysis and topographic surveys to quantify the recovery of a beach ridge plain in West Aceh, Sumatra, Indonesia over a time span of 9 years after the 2004 Indian Ocean tsunami, and determine the effects of large-scale events on beach ridge morphologies and growth patterns. We will compare the obtained morphological data to previous paleoseismic studies of the coastal sediments in this area (Monecke et al., 2008) and discuss if beach ridge morphologies are suitable to reconstruct past earthquake histories.

show extensive dune development on top of ridges; such strandplains are sometimes identified as dune or foredune ridges (e.g. Wells and Goff, 2007; Masselink et al., 2011) or might be included under the term beach ridge plain (e.g. Otvos, 2000; Bristow and Pucillo, 2006). If aeolian processes are secondary, beach ridge evolution is primarily a result of swash processes under varying wave energy conditions (Taylor and Stone, 1996). A low shoreface gradient as well as an abundance of sediment are favourable for progradation of beach ridges (Taylor and Stone, 1996; Bristow and Pucillo, 2006) with sediment supply being the most important factor controlling the spatial and temporal variation in beach ridge development (Anthony, 1995). Sediment supply to the littoral environment can increase significantly during extreme climate events (Goy et al., 2003; Shafer Rogers et al., 2004) or the mobilization of sediment as a result of earthquake shaking (Wells and Goff, 2007). Geometries of beach ridge systems have successfully been used to reconstruct sea and lake level histories on millennial to decadal scales (Tanner, 1995; Thompson and Baedke, 1995; Engels and Roberts, 2005; Storms and Kroonenberg, 2007). Rates of beach ridge progradation are commonly determined by optical luminescence dating of sandy ridge deposits (e.g. Ballarini et al., 2003; Bristow and Pucillo, 2006) but radiocarbon dating of organic material from within beach ridges (e.g. Goy et al., 2003) or from inter-ridge swales (e.g. Thompson, 1992) has been used as well. In a few cases beach ridge patterns have been shown to reveal earthquake cycles. Bookhagen et al. (2006) interpret a sequence of uplifted beach berms in South-Central Chile as paleoshorelines that were progressively exposed during earthquakes along the NazcaSouth America plate interface. Briggs et al. (2008) document a set of seaward climbing beach berm crests on Nias Island in West Sumatra, which might respond to slow interseismic subsidence along this part of the Sunda trench subduction zone. Mobilization of large amounts of sediments produced during earthquake shaking is suggested as the driving force for the formation of dune ridges in New Zealand (Wells and Goff, 2007). A link of beach ridge formation and earthquake activity is tentatively suggested for a strandplain in eastern Japan, yet, the complex interplay of seismic, volcanogenic and climatically-driven processes needs more thorough investigation (Goff and Sugawara, 2014).

1.1. The morphology of beach ridge plains

1.2. Land level changes during subduction earthquakes

Many parts of the Sumatran coast consist of beach ridge plains, where shore-parallel sand ridges alternate with marshy swales, typical for a prograding coastline. Although their geometries look quite regular, large magnitude earthquakes along the Sunda trench cause vertical ground displacement and tsunamis of devastating force which can modify coastlines significantly, as evidenced by the December 26, 2004 Sumatra-Andaman earthquake and ensuing Indian Ocean tsunami (Liew et al., 2010). In Aceh, the northernmost province of Sumatra, Indonesia, a retreat of the coastline of several tens of meters was observed and can be attributed to coseismic subsidence, tsunami scouring and sediment redistribution (Jaffe et al., 2006; Meltzner et al., 2006; Meilianda et al., 2010). More recent satellite images, however, show that in many places the shoreline has built rapidly seaward forming a new prominent beach ridge since the 2004 event (Liew et al., 2010). The formation of beach ridge coastlines and possible modifications by large-scale events is still a matter of debate (see Tamura, 2012; for a review). A beach ridge plain is characterized by sets of sandy ridges separated by low-lying swales that run parallel or subparallel to the coast for several kilometres. Individual ridges mark former positions of the shoreline with older ridges located further inland while new ones build progressively seaward. Prograding coastlines experiencing significant wind transport may

It can be assumed that, in seismically active areas, growth rates of beach ridge plains are severely affected by coseismic land level changes, which affect the local sea level and shoreface gradient. Large subduction earthquakes cause a characteristic pattern of vertical ground displacement: Uplift up to a few meters occurs closer to the trench often visible by emergence of offshore islands, while subsidence dominates in the forearc region causing the drowning of coastlines (e.g. Plafker, 1972). Significant land level changes continue after the event and comprise afterslip in the months following the earthquake, mantle relaxation over decades after the event and relocking of the fault thereafter (Wang et al., 2012). The latter two processes are generally opposite in direction to coseismic slip; areas experiencing coseismic uplift are subsiding whereas coseismically subsiding areas are uplifted (e.g. Savage and Plafker, 1991; Natawidjaja et al., 2007; Suito and Freymueller, 2009). Continuous GPS measurements along many active margins (but not in Aceh), over the past 20 years have allowed the quantification of co-, post- and interseismic deformation and the development of sophisticated earthquake dislocation models that predict the land level changes in subduction zones over one earthquake cycle (for a review see Wang et al., 2012). Since the 2004 Sumatra-Andaman earthquake and ensuing tsunami caused relative sea level rise along the coastline of West Aceh, in addition

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

to nearshore erosion and significant sediment redistribution, we anticipate that it left a distinct imprint on beach ridge stratigraphy. 2. Regional setting 2.1. Seismotectonic setting The Sunda trench subduction zone parallels the west coast of Aceh at a distance of 250e300 km (Fig. 1). Large earthquakes have been generated along this megathrust including the recent Mw 9.2 December 26, 2004 Sumatra-Andaman earthquake with a rupture length of 1600 km and slip locally exceeding 20 m (Chlieh et al., 2007). A combination of satellite image analysis, GPS campaign data and coral measurements indicate coseismic uplift of up to 2 m on offshore islands close to the trench and subsidence of 0.5e1 m along the west coast of Aceh (Fig. 1, Meltzner et al., 2006; Subarya et al., 2006). Postseismic measurements from a permanent GPS station (Ujung Muloh, UMLH, in Fig. 1) of the Sumatra GPS Array (SuGAR) indicate continued horizontal displacement of the west coast of Aceh but nearly no vertical displacement or slight subsidence in the year following the earthquake, which can be attributed to afterslip (Shearer and Bürgmann, 2010). Since late 2005 the coastline experiences uplift of 27 mm/year (processed data from UMLH station made available by Jeff Freymueller, University of Alaska, Fairbanks, 2014), probably the result of postseismic mantle

3

relaxation. Viscoelastic mantle relaxation models suggest decreasing rates of postseismic uplift in Aceh over the next few decades amounting to a total of 50 cm for a period of 60 years after the Sumatra-Andaman earthquake (Shearer and Bürgmann, 2010). The historic seismic record of Indonesia revealed no major earthquake or tsunami affecting the west coast of Aceh for the last 400 years (Newcomb and McCann, 1987; Hamzah et al., 2000). One exception is the A.D. 1907 earthquake that generated a tsunami devastating the coastal areas of offshore Simeulue Island (Fig. 1) but reaching only minimum heights along the Acehnese mainland (McAdoo et al., 2006). More recent studies of older historical documents point to tsunami inundation in A.D. 1349 and around A.D. 1000 (Nurjanah, 2013). Geologic evidence of past seismic activity is recorded in the growth pattern of corals from northern Simeulue Island (Fig. 1), which lies within the 2004 rupture area (Meltzner et al., 2010). Here, abrupt coseismic land level changes occurred during an earthquake cluster between A.D. 1390e1455 with considerable greater uplift than in 2004. The sediments of a beach ridge plain 15 km northwest of Meulaboh in West Aceh yielded evidence of past tsunami inundation through buried sand sheets deposited soon after A.D. 780e990 and A.D. 1290e1400 (Monecke et al., 2008). The combined historical and geologic record from Aceh suggests that the last predecessor of the 2004 SumatraAndaman earthquake and tsunami occurred about 600 years ago, which has been confirmed along other coastlines in the Indian Ocean (e.g. Jankaew et al., 2008). 2.2. Geomorphology of West Acehnese coastline

Fig. 1. Northern Sunda Margin and vicinity. Gray patch marks rupture area of December 2004, Mw ¼ 9.2 Sumatra-Andaman earthquake after Chlieh et al., 2007. Coseismic land level changes during 2004 earthquake are from satellite image observations, uplifted corals and GPS campaign data after Subarya et al. (2006). Closest tide gauge station is located in Meulaboh (Me). Coral observations by Meltzner et al. (2010) were carried out on Simeulue Island (Si). Triangles mark permanent GPS stations in Aceh installed in 2005.

Here, we are revisiting the previously studied beach ridge plain in West Aceh (Monecke et al., 2008), 15 km northwest of Meulaboh (Figs. 1 and 2). The study area is part of the Meulaboh Embayment, an extensive low-lying coastal plain southwest of the Barisan Mountains, which run the length of Sumatra. The sediments of the Meulaboh Embayment consist of a few hundred meter thick sequence of Plio-Pleistocene clastic sedimentary rocks and Quaternary gravels, sands and clays deposited in a fluvial to coastal environment (Cameron, 1983). Geophysical and borehole data suggest a large prograding delta complex with cyclic sedimentation over the last 2.6 Ma with the sediments of the beach ridge plain forming the top few meters of a larger sand body (Finger, 2007). The Woyla River (Fig. 2), as well as the Meulaboh River to the south, are major rivers in this area and carry abundant sediment from the Barisan Mountains to the coastline. Uplift of the Barisan Mountains might have started as early as the Miocene and the presence of river terraces as well as deeply incised river valleys suggest that uplift continues at present (Cameron, 1983). Even though we have no direct measurement of sediment yield of the two rivers, uplift combined with a tropical climate point towards a large sediment supply to the coastal zone. The studied beach ridge plain extends approximately 2 km inland and stretches along the coast for 10 km between the Woyla River in the Northwest and a raised reef structure that forms a small peninsula in the Southeast (Fig. 2a). The overall topography of the study area is low with individual beach ridge crests reaching no more than 2.5 m height above sea level (Fig. 2b). The ridges run subparallel to the coast, curve towards the raised coral reef structure in the Southeast, and get wider and less pronounced towards the Woyla River floodplain in the Northwest (Fig. 2a). Sandy ridges are separated by swales where peaty marsh deposits accumulate and can be dissected by artificial or natural shore-normal channels. The coastal road as well as several houses are located on a higher ridge in about 640 m distance to the shoreline (Fig. 2). Large parts of the beach ridge plain are covered by dense swamp vegetation and inland access is only possible along a few man-made trails.

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

4

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

Fig. 2. a) Overview of study area and beach ridge plain north of Meulaboh. Beach ridges run sub-parallel to the shore and can be followed laterally over several kilometers. Intervening swales are submerged in this April 2005 image due to coseimic subsidence. Tsunami wave heights reached 9e14 m in this area and the limit of tsunami inundation is marked by the extent of saltwater affected trees showing up in gray. Observed extent of 2004 tsunami deposit and buried sand sheets related to earlier tsunami inundation (Units B and C) are from Monecke et al. (2008). Photographs were made available by the SIM Center of the Aceh and Nias Rehabilitation and Reconstruction Board (BRR). b) Simplified cross section of beach ridge plain indicating the long-term seaward progradation of beach ridges based on radiocarbon dating of swale deposits, the presence of outsized ridges and the observed extent of tsunami sand sheets after Monecke et al. (2008).

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

The sequence of ridges builds progressively seaward at a rate that can be estimated from the age of deposits accumulating on the beach ridge plain. When a new beach ridge starts to form, the area landward of it becomes protected from wave action and organicrich sediment starts to accumulate. The age of the oldest deposit in a swale can therefore give an estimate of the beach ridge immediately before it. The oldest deposits in swales in about 1800 m and 950 m distance to the current shore have been dated to A.D. 780e990 and A.D. 1290e1400, respectively (Fig. 2), thus indicate average progradation rates of 1.5e1.8 m/year and 1.3e1.6 m/year (Monecke et al., 2008). Even though the deepest swale sediments might not have been deposited exactly synchronously to the adjacent beach ridge, our ages are consistent and match progradation rates reported from beach ridge plains elsewhere (e.g. Anthony, 1995; Tanner, 1995; Bristow and Pucillo, 2006; Brooke et al., 2008). In low-lying areas of the beach ridge plain, the 2004 tsunami deposited a sand sheet that can be followed up to 1800 m inland (Fig. 2, Monecke et al., 2008). While a continuous, up to 50 cm thick sand sheet can be found in swales up to 600 m inland, the deposit becomes patchy and thins to a few millimeters farther away from the shoreline. Sediment cores taken from swales between beach ridges revealed two older sand layers intercalated with peaty marsh deposits (Unit B and C in Fig. 2, Monecke et al., 2008). In keeping with the progressive seaward migration of the shoreline, the older sand layer (Unit C) can be found farther inland than the younger Unit B. These sand sheets deposited soon after A.D.

5

780e990 and A.D. 1290e1400, were probably laid down by earlier tsunamis (Monecke et al., 2008). 2.3. Hydrodynamic setting The rebuilding of a shoreline after extreme events is largely dependent on the prevailing hydrodynamic conditions (e.g. Yu et al., 2013). The coastline of West Aceh can be characterized as a wave dominated environment with the wave climate being controlled by the East Monsoon from October to March and the more energetic West Monsoon from April to September. We have no directly measured wave data, but a wave climate model indicates calmer conditions with a significant wave height (Hs) of 0.95 m during the East Monsoon and greater wave heights reaching Hs ¼ 1.2 m during the West Monsoon (wave data modelled in 7 m water depth after de Graaff, 2007). Waves approach the shoreline at a very low angle from a WNW to ESE direction, resulting in minor and variable longshore sediment transport. A tide gauge station installed as part of Indonesia's Tsunami Early Warning System in nearby Meulaboh (Fig. 1) indicates a microtidal environment with a semidiurnal tide and a maximum tidal range of 0.5e0.6 m during € ne et al., 2011; quality-controlled data available spring tides (Scho from the University of Hawaii Sea Level Center, UHSLC). The beach can be classified as intermediate and has an average foreshore slope of tanb ¼ 0.07. The foreshore and beach environment is composed of well sorted medium sand and comprises predominantly silicate grains as well as lesser amounts of heavy minerals. Medium

Fig. 3. Detail of Fig. 2 showing coastal development since 2002 and paths of auto-level surveys in 2009, 2012 and 2013. a) Prior to the 2004 earthquake and tsunami a wide beach is visible. b) Coseismic subsidence and tsunami inundation cause large coastal retreat in 2004. Note submergence of swales and low-lying areas. Incipient beach formation is visible 4 months after 2004 earthquake. c) Rapid coastal progradation causes the formation of a new wide beach ridge. d) Renewed coastal retreat. Note progressive emergence of swales and low-lying areas in c) and d) due to postseismic uplift.

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

6

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

Table 1 Uncertainties related to topographic surveys and spatial imagery analysis. RMS is Root-Mean-Square error. See text for further explanation. Topographic surveys Date of survey

3/10/2009 8/31/2012 6/12/2013

Horizontal error

Vertical error

Measurement error (m)

Instrument error (m)

Measurement error (m)

Run-up variation (m)

RMS error (m)

5 10 15

0.0007 0.002 0.002

0.1 0.02 0.005

0.52 0.66 1.45

0.53 0.66 1.45

Spatial imagery analysis Source of image

Date of image

Monsoon

Cell size (m)

Georeferencing offset (m)

Tidal variation (m)

Run-up variation (m)

RMS error (m)

Pleiades WorldView-2 WorldView-2 Quickbird Quickbird Ikonos Aerial Photos (BRR) Quickbird

6/28/2013 1/31/2011 04/15/2011 12/08/2009 12/02/2009 06/26/2006 04/05/2005 09/19/2002

SW NE SW NE NE SW SW SW

0.5 0.5 0.5 0.6 0.6 0.8 0.3 0.6

5.5 3.7 4 1.4 1.8 3.5 0 4.5

8 8 8 8 8 8 8 8

13.2 13.2 13.2 13.2 13.2 13.2 13.2 13.2

16.4 15.9 16.0 15.5 15.6 15.8 15.4 16.1

resolution bathymetric data indicates an upper shoreface slope of 1/100 with a depth of 5 m being reached in 500 m distance to the coast (personal communication by Widjo Kongko, Coastal Dynamic Research Institute, BPPD, Yogyakarta, Indonesia). Further offshore the slope decreases to 1/200 with 10 m water depth being reached in about 1500 m distance to the shoreline. 3. Methods In order to quantify the short-term coastal evolution since the 2004 Sumatra-Andaman earthquake and Indian Ocean tsunami and to unravel the large-scale architecture of the beach ridge plain, we conducted topographic surveys and analysed spatial imagery of the coastal marshes in West Aceh. 3.1. Topographic surveys Three topographic surveys were conducted in March 2009, August 2012 and June 2013 by using an automatic level suitable for a remote, densely vegetated environment allowing height, distance and angle readings (Fig. 3). Distances between surveyed points were cross-checked with a range finder and a handheld GPS and later evaluated using Geographic Information Systems (GIS) software. The surveys in 2012 and 2013 were carried out along transects running perpendicular from the shoreline up to 700 m inland, ending behind a prominent beach ridge on top of which the coastal road and several houses are built (Figs. 2 and 3). In 2009, the survey extended up to 1800 m inland before dense swamp vegetation hindered any further progress. While all three transects are located in the same general area, individual survey points differ from year to year because of changing vegetation patterns, landuse and flooding. However, since the beach ridges form a regular shoreparallel pattern, adjacent shore-normal paths should capture the same sequence of ridges and swales. Because of the lack of benchmarks in this area, vertical and horizontal reference points had to be established, which would allow comparison of the three surveys. The vertical reference for individual transects is the high water line (HWL), the furthest landward extent of the last high tide, which was corrected relative to the tidal pattern as recorded at the tide gauge station in Meulaboh (quality-controlled data available from the University of Hawaii Sea Level Center, UHSLC). The position of the coastal road and the deepest spot in the most seaward swale serve as horizontal reference points.

An error analysis was performed for each transect to account for instrument errors and measurement errors (Table 1). The largest vertical error in our topographic surveys results from the interpretation of the high water line. The HWL can be approximated by markers left by the previous high tide like driftwood or the boundary between wet and dry sand (e.g. Pajak and Leatherman, 2002; Boak and Turner, 2005). The position of the HWL on the beach is a result of the tidal elevation as well as wave run-up, the upward displacement of the shoreline due to wave set-up and swash (Ruggiero et al., 2001; L. J. Moore et al., 2006). Wave run-up is largely dependent on beach slope and wave height (Masselink et al., 2011). Using the relationship of Holman (1986), the extreme wave run-up R2%, the run-up height exceeded by 2% of all run-up events, can be determined. Although we do not have measured wave data for the times of our surveys, we calculated R2% for the range of slopes we measured in 2009, 2012 and 2013 (0.04, 0.05 and 0.11, respectively) and an average wave height of Hs ¼ 1.08 m (from modelled wave data from de Graaff, 2007). R2% varies between 0.52 m and 1.45 m (Table 1), thus indicating that run-up could have caused significant displacements of HWL and therefore, of all points on our transects. 3.2. Spatial imagery analysis Eight georeferenced and orthorectified satellite images and a set of aerial images covering the time span from 2002 to 2013 were analysed to quantify horizontal shoreline change along an 18 km long stretch of the western coast of Aceh (Fig. 4, Table 1). On all imagery we digitized the wet/dry line, a distinct tonal contrast between wet and dry sand and a good approximation of the high water line (Moore, 2000; Pajak and Leatherman, 2002). We then used the Digital Shoreline Analysis System (DSAS) developed by the USGS (Thieler et al., 2009) to quantify coastal retreat or progradation. A baseline was defined seaward of all shorelines and 375 transects were cast perpendicular to the baseline across all digitized shorelines at 50 m intervals (Fig. 4). Along each transect the shoreline change between subsequent shorelines was calculated and averaged over all transects. Shoreline positioning errors are the result of the resolution of the images, georeferencing offsets, and the tidal range, since the time of acquisition of the satellite images is not known and/or the tidal record is not available for images taken before 2008 (Table 1). In addition, the variation of run-up contributes to uncertainties in our analysis as outlined above. The resolution of images is highest

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

Fig. 4. Shoreline change analysis. Shorelines are digitized on satellite images from September 2002 to June 2013 using the wet/dry line as illustrated for the 2013 shoreline in this 2013 satellite image (violet line). All shorelines were intersected by transects set perpendicular to a baseline using the Digital Shoreline Analysis System (DSAS) developed by the USGS (Thieler et al., 2009). The baseline extends for 18 km along the West Acehnese coast; shoreline change was calculated along 375 transects. Depicted shorelines show characteristic pattern of coastal evolution in West Aceh since 2002. Note the large retreat following the 2004 earthquake and tsunami (2002 vs. 2005 shore), followed by rapid regrowth until 2011 and a renewed retreat from 2011 to 2013. See text and Table 1 for uncertainties in shoreline position and Table 2 for absolute values of shoreline change.

on a set of aerial photographs with a ground cell resolution of 30 cm and lowest on 2013 satellite images with a resolution of 80 cm (Table 1). The original georeferencing was cross-checked by comparing the position of stable control points identifiable on all images, such as mosques. All points were within less than ±5.5 m of each other. The variation of high tide excursions measured over one month is 0.56 m; assuming an average slope of tanb ¼ 0.07, this leads to a horizontal shoreline variation of ±8 m. The largest error is associated with variable run-up (average R2% ¼ 0.92 m) leading to an error of ±13.2 m in shoreline position. The sum of all errors (Root-Mean-Square, RMS) of individual shoreline positions is less than ± 16.4 m (Table 1). 4. Results 4.1. Results of topographic surveys Topographic survey data from March 2009, August 2012 and June 2013 are illustrated in Fig. 5. All three transects show the characteristic alternation of ridges and swales and the overall low morphology of the beach ridge plain with maximum elevations not exceeding 1.7 m above the high water mark as surveyed on that day. Some ridges display a typical asymmetrical shape with a steep seaward face and a more gently sloping landward face. The most seaward ridge that formed since the dramatic retreat of the coast in December 2004 is the most prominent ridge in all three transects. It is between 106 and 163 m wide with its crest 0.9, 1.5 and 1.7 m above the respective high water mark in 2009, 2012 and 2013 (Fig. 5). In 170e300 m distance to the shoreline beach ridges are narrower and distinctly lower with widths ranging from 20 to 40 m and crest heights 0.8e1.3 m below the crest of the prominent most

7

seaward ridge. Interspersed swales reach depths of 0.4e1.0 m relative to adjacent ridge crests. Further inland, between 300 and 640 m distance to the shoreline, the characteristic morphology of ridges and swales is less pronounced, partly because flooded areas obstructed a direct line of survey and less data points were collected here. However, variations in elevation are small and in the same range as for the sequence of lower ridges described above. The most landward ridge surveyed by all three transects in 640 m distance to the coastline is another prominent ridge along which the coastal road is built (see Fig. 2). This ridge is about 50 m wide and stands 0.5e0.9 m higher than the set of lower ridges seaward of it (Fig. 5). When comparing the three transects it has to be taken into account that the paths of the three transects are not identical and that the vertical reference point is poorly constrained due to varying wave run-up. Nevertheless, the sequence of transects indicates a general trend of relative sea level fall or coastal uplift. This trend is confirmed by the deepening of nearshore swales that drain towards the Indian Ocean and thus, respond to relative sea level changes. The three most seaward swales that were surveyed in close proximity in the first two years are on average 0.17 m deeper in August 2012 compared to March 2009 (Fig. 5). Field observations in 2012 and 2013 revealed significant coastal retreat resulting in the 2013 shoreline being located 44 m landward of the 2009 shore (Fig. 5). 4.2. Horizontal shoreline change analysis Fig. 4 and Table 2 summarize the shoreline change in West Aceh as observed on spatial imagery taken between 2002 and 2013. Comparison of pre- and post-tsunami imagery (September 2002 versus April 2005) indicate an average coastal retreat of 110 m over the investigated 18 km long stretch of coastline in the course of the December 2004 earthquake and tsunami. This is a minimum estimate, since the coast had already started to accrete in places by the time the spatial imagery was taken in April 2005. The retreat was largest where natural creeks and drainage channels discharge into the ocean. The tsunami bored out these river outlets as protective sand barriers were removed during tsunami inundation (compare Fig. 3a and b). Following the retreat, the coastline grew back rapidly at a rate of 18 m per year with an average progradation of 22 m between April 2005 and June 2006. In this phase, the previously widened creeks and channel mouths were partially closed again resulting in the largest progradation rates in these areas. Coastal growth slowed to 9 m per year from June 2006 to December 2009 resulting in an additional advance of 30 m. Between December 2009 and April 2011 the coastline seemed to have stabilized and experienced only minor changes (Fig. 3c). At this point the coastline was still on average 45 m from its pre-tsunami position. Comparison of the 2011 images to our latest set of images from June 2013 reveals a renewed retreat of the coastline of an average 41 m (Fig. 3d), which closely matches our field observations. The 2013 shoreline is located on average 86 m landward of its pre-December 2004 position. 5. Discussion 5.1. Beach ridge formation following the December 2004 earthquake and tsunami Coastal development in West Aceh in response to co- and postseismic land level changes, tsunami inundation and sediment redistribution following the December 2004 earthquake shows a characteristic pattern (Fig. 6). Before the earthquake a wide beach can be identified on satellite images (see Fig. 3a). At this stage, the

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

8

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

Fig. 5. Topographic survey data. Transects show alternating ridges and swales with the most seaward ridge that formed since December 2004, being the most prominent one. Another prominent ridge on which the coastal road is located, occurs in 640 m distance to the shoreline. Later surveys in 2012 and 2013 indicate uplift and incision of swales and retreat of the coastline. Black symbols mark high water line with error related to varying wave run-up (see Table 1 and text for further explanation). Elevation of transects is corrected relative to tide gauge data from Meulaboh for the time of survey (quality-controlled tidal record available through the University of Hawaii Sea Level Center, UHSLC). Note vertical exaggeration.

Table 2 Results of Shoreline Change Analysis using the Digital Shoreline Analysis system (DSAS, Thieler et al., 2009). Shoreline change was calculated along 375 transects cast perpendicular to an 18 km long stretch of the coastline in West Aceh. See Table 1 and text for uncertainties in shoreline position. Period

9/19/2002 e 4/5/2005 4/5/2005 e 6/26/2006 6/26/2006 e 12/2/2009 6/26/2006 e 12/8/2009 12/2/2009 e 1/31/2011 12/2/2009 e 4/15/2011 12/8/2009 e 1/31/2011 1/31/2011 e 6/28/2013 4/15/2011 e 6/28/2013 9/19/2002 e 6/28/2013

Average shoreline change (m)

Average shoreline change rate (m/yr)

With outlets

Without outlets

With outlets

Without outlets

134 55 27 38 4 1 9 36 47 86

110 22 26 n/a 4 1 n/a 38 47 87

n/a 45 8 11 3 1 7 15 20 n/a

n/a 18 8 n/a 4 1 n/a 16 20 n/a

Transects used

32e371 19e371 120e372 32e127 120e314 234e375 33e127 33e314 234e346 32e346

typical morphology of ridges and swales is barely visible, since large parts of the beach ridge plain are above sea level including depressions and shore-parallel swales (Figs. 3a and 6a). While we do not have topographic survey data from before the earthquake, we assume that the height of the most seaward beach ridge was in the same range as of beach ridges further inland reflecting a relatively stable sea level during the late interseismic interval preceding the earthquake. Coseismic subsidence of 0.5e1 m in December 2004 (Meltzner et al., 2006; Subarya et al., 2006) caused large areas of the beach ridge plain to be submerged with widespread flooding of shoreparallel swales (Figs. 3b and 6b). Subsidence as well as tsunami inundation and erosion resulted in a nearly complete removal of the most seaward beach ridge and the displacement of the coastline up to 200 m inland. The mobilized sediment was redistributed and either deposited in low-lying areas of the coastal marshes (Monecke et al., 2008) or transported offshore with the backwash

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

9

Fig. 6. Coastal Development Model. a) Most parts of the beach ridge plain are above sea level before the 2004 earthquake and tsunami. Only deepest points of swales are submerged. b) Shortly after the 2004 earthquake and tsunami large areas of the beach ridge plain are submerged as a result of coseismic subsidence. The coastline is displaced more than 200 m inland. Eroded material is transported onshore and offshore. c) In the months and years after the event a new higher beach ridge forms corresponding to the higher relative sea level. d) Postseismic uplift causes emergence of coastal area and readjustment of coastal profile in years to decades after the earthquake.

of the tsunami. Modelling of tsunami inundation and sediment transport along a cross-shore profile measured about 50 km north of our study area, and comparable in shape to bathymetric data from West Aceh, indicates highest erosion near the original shoreline, minor deposition onshore and largest sediment accumulation offshore in 500e1000 m distance to the coast (Apotsos et al., 2011). These simulations further suggest the formation of a large bar at the seaward end of the tsunami backwash in about 1000 m distance to the shore and water depths between 10 and 12 m, where the backwash collides with the onrush of the next wave (Apotsos et al., 2011). This is consistent with bathymetric data from offshore surveys conducted in the months after the 2004 tsunami (Apotsos et al., 2011; personal communication with Peter Ruggiero, Oregon State University). Similar coastal profile changes including offshore sediment transport, beach erosion and the formation of a nearshore bar can be observed after large storm events (Masselink et al., 2011); however, they will affect a smaller part and shallower depths of the cross-shore profile. In the years following the December 2004 earthquake, the coastline steadily prograded, more rapidly in the first few months and then at moderate rates until reaching an equilibrium some time between 2009 and 2011 (Figs. 3c and 6c). While falling short of its pre-tsunami position by 45 m, a new wide beach ridge had formed. Relatively quick beach accretion after extreme events has been observed along the eastern coast of Japan following the 2011 Tohoku-oki tsunami (e.g. Tappin et al., 2012), after the 2004 Indian Ocean tsunami in Aceh, Indonesia (e.g. Liew et al., 2010; Meilianda et al., 2010), and in Thailand (Choowong et al., 2009) as well as after large storms on Hongkong Island, Southern China (Yu et al., 2013) and Florida, USA (Wang et al., 2006). Beach accretion in these cases

can be attributed to the establishment of regular coastal processes, when low to moderate energy conditions cause the reworking of offshore bars, shoreward directed sediment transport and beach accretion (Masselink et al., 2011). Bathymetric data obtained two years after the tsunami about 50 km north of our study area at the same location as discussed above, indicate that the previously described bar in about 1000 m distance to the shore had been smoothed out (personal communication with Peter Ruggiero, Oregon State University). A high energy environment, which is also prevailing along the western Acehnese coast, seems to be favourable for a quick reworking of offshore transported sediment and beach recovery (Yu et al., 2013). The newly formed beach ridge in our study area is 0.8e1.3 m higher than beach ridges further inland (Fig. 5). The height of ridges and the presence of outsized berms can either be a reflection of dune development on top of beach ridges or changes in sea level, if ridges are swash-built and aeolian processes can be excluded (Tamura, 2012). Coastal dunes usually develop on progradational, sand-rich coasts with energetic wind conditions (Masselink et al., 2011). While specially adapted pioneering vegetation colonises and stabilizes wind-transported sand, dense backbeach vegetation that is often found in the tropics, hinders the development of coastal dunes (Masselink et al., 2011). The presence of tall outsized foredune ridges on strandplains in Australia has been explained by low rates of progradation that allow more time for aeolian sand transport from the beach to the foredune and the construction of higher foredune ridges (Bristow and Pucillo, 2006; Brooke et al., 2008). We found that aeolian processes probably only play a minor role in the construction of ridges in West Aceh. The seaward side of the

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

10

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

beach is essentially bare of vegetation hence, the trapping of sediment is limited. Further landward large dune development is hindered in places by dense swamp vegetation. We have experienced strong winds and witnessed aeolian sediment transport during fieldwork in West Aceh (e.g. in June 2013 during a relatively powerful West Monsoon), however, we have not observed the longterm establishment of linear foredune ridges. Our topographic survey data display little variation in the height of the most seaward beach ridge from 2009 to 2013 (Fig. 5) thus, confirming that dune growth on ridges is limited. During an earlier fieldwork in March 2006, we were able to analyse the internal structure of the newly forming beach ridge just south of our study area, where a trench had been cut to excavate sand for the reconstruction efforts in West Aceh. The outcrop showed 1.4 m of well-stratified sand with heavy mineral laminations and low angle cross stratification confirming the prevalence of swash processes in beach ridge formation. In conclusion, the difference in height of 0.8e1.3 m between the most seaward beach ridge and the sequence of ridges further inland most likely reflects the higher relative sea level induced by coseismic subsidence in 2004. Such an interpretation is supported by the fact that the height difference matches the amount of coseismic subsidence of 0.5e1.0 m of the northern Sumatran coastline during the December 2004 earthquake (Meltzner et al., 2006; Subarya et al., 2006). This observation confirms our initial assumption that land level changes associated with large earthquakes along the Sunda trench leave a distinct imprint on beach ridge morphology. While other coastlines have nearly completely recovered to their pre-2004 position, e.g. in Thailand where no land level changes occurred (Choowong et al., 2009), the West Acehnese coast remained 45 m landward from its pre-2004 position. Similar incomplete coastal recovery after the 2004 tsunami has been observed in northern Aceh, where only 60% of the sediment initially lost to the tsunami was restored in the months following the event (Meilianda et al., 2010). Migration patterns of shorelines and associated coastal deposition is a function of sediment supply, eustatic sea-level changes and subsidence (Helland-Hansen and Hampson, 2009). In Aceh, coseismic subsidence (Meltzner et al., 2006; Subarya et al., 2006) caused significant changes in relative sea level resulting in a rapid landward shift of the shoreline. As described above, sediment returned quickly to the shoreline in the months following the earthquake, however, it was not sufficient to fill the excess space created by coseismic subsidence. In addition, it can be assumed that on- and offshore sediment redistribution in 2004 caused significant sediment shortage within the coastal system. The 2004 tsunami mobilized huge quantities of sediment from the nearshore environment (Apotsos et al., 2011), part of which was deposited within the coastal marshes as an extensive sand sheet (Monecke et al., 2008). Post-tsunami bathymetric surveys as well as numerical modelling of sediment transport under tsunamis further indicate that large amounts of sediment were moved into depths greater than 10 m (Apotsos et al., 2011), which are not regularly affected by wave action, thus remain at least temporarily lost from the coastal system. The combination of a higher relative sea level and reduced sediment supply is the most likely cause for the incomplete recovery of the coastline in West Aceh. 5.2. Preservation potential of seismically modified beach ridge morphology Recent topographic survey data from August 2012 and June 2013 as well as satellite images from June 2013 indicate a renewed retreat of the coastline (Figs. 3 and 5). This poses the question of the preservation potential of beach ridge morphology altered by extreme events. Since all of the recent data was obtained during the

more energetic West Monsoon, we cannot exclude that seasonal high energy conditions have caused the most recent retreat of the shoreline. However, another factor controlling the coastal migration pattern is postseismic uplift (Fig. 6d). Permanent GPS measurements in northern Aceh indicate rapid postseismic uplift starting in late 2005, which has amounted to 22 cm by 2013 (UMLH station in Fig. 1, processed data made available by Jeff Freymueller, University of Alaska, Fairbanks, 2014). This trend is visible in our topographic surveys (Fig. 5) and in satellite images, showing the drying up of swales and depressions (2005 vs. 2011 and 2013 images in Fig. 3). Relative sea level fall causes erosion of coastal sediment, the establishment of a lower depositional profile and progradation, if sediment can be readily made available (HellandHansen and Hampson, 2009). Storms and Kroonenberg (2007) observe that different beach ridge systems along the Caspian See experiencing the same amount of rapid sea level fall respond very differently. Only where sediment is readily available the coastline is prograding. At a site in northern Aceh that can be described as a “sediment-poor” environment, Meilianda et al. (2010) noticed renewed shoreline erosion starting in 2006 after a brief phase of accretion immediately following the tsunami. The authors suggest that the growth pattern of shorelines in Aceh is largely controlled by the inherited coastal morphology and sediment availibility, thus, determining if the coastline experiences long-term post-tsunami erosion or accretion (Meilianda et al., 2010). We suggest that the observed sediment shortage within the coastal system in West Aceh following tsunami inundation results not only in an incomplete recovery of the shoreline but also in limited availability of sediment for the adjustment of the cross-shore profile in response to rapid postseismic relative sea level fall. Postseismic uplift is predicted to prevail over the next 50 years and could amount up to an additional 30 cm (Shearer and Bürgmann, 2010). Uplift rates, however, will most likely decrease over time as has been observed e.g. for areas that subsided during the 1964 Alaska earthquake and experienced postseismic uplift thereafter (Suito and Freymueller, 2009). We expect that decreasing land level changes will result in stabilization of the cross-shore profile and establishment of equilibrium conditions on the West Acehnese coastline within the next few decades. Since uplift in the Barisan Mountains is ongoing, sediment supply to the coastal zone will likely be high and replenish the sediment lost during tsunami inundation. Once equilibrium is established, the coastline will most likely prograde, as is evident in the long-term stratigraphy of the beach ridge plain. An indication of long-term preservation of outsized beach ridges forming after large subduction earthquakes can be found within older beach ridges further inland. The coastal road and houses alongside of it are built in large parts on a prominent ridge in 640 m distance to the shoreline (see Fig. 2). This ridge has also been identified in our topographic surveys and is 0.5e0.9 m higher than beach ridges further seaward (Fig. 5). While road construction most likely involved some modification of the ridge due to levelling, compaction and filling, we suggest that the initial prominent morphology was crucial for choosing the location of the village and the path of the coastal road in the first place. We propose that, similar to the ridge that formed in the aftermath of the 2004 earthquake, this prominent ridge formed in response to coseismic subsidence of the last predecessor of the 2004 earthquake about 600 years ago. Considering the long-term progradation of the beach ridge plain, the coastline at that time was probably located in the vicinity of the prominent ridge in 640 m distance to the shoreline. Coseismic land level changes during a 2004-like rupture would have caused subsidence in Aceh and triggered the formation of an outsized ridge in the aftermath of the earthquake. Based on coral evidence, Meltzner et al. (2010) suggest that such significant

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

land-level changes must have occurred during an earthquake cluster at the south end of the 2004 Sumatra-Andaman rupture between A.D. 1390e1455, which would have also affected the coastline in Aceh. Further evidence comes in form of a sand sheet found in swales landward of the prominent beach ridge and most likely deposited during tsunami inundation shortly after A.D. 1290e1400 (Unit B in Fig. 2, Monecke et al., 2008). Since the thickness of tsunami deposits and thus, their preservation potential is highest in low-lying areas close to the original shore (e.g. A. Moore et al., 2006), the sand sheet is most likely linked to the formation of the prominent ridge immediately before it. An additional tsunami sand sheet deposited shortly after A.D. 780e990 was found within beach ridges further inland (Unit C in Fig. 2; Monecke et al., 2008). However, here we don't have clear evidence of an outsized ridge paired with this tsunami deposit, partly because survey data was limited in the densely vegetated and partially flooded inland areas (Fig. 3). Based on preliminary observations of beach ridge morphologies in eastern Japan, Goff and Sugawara (2014) have suggested a dominantly seismic origin of beach ridges on the coastal plain of West Aceh implying the occurrence of more earthquakes and tsunamis than initially proposed by Monecke et al. (2008). We find that beach ridge formation in Aceh is primarily triggered by abundant sediment supply to the coastal zone from an uplifting source area causing the longterm progradational trend of the shoreline. A seismic origin is only inferred for outsized beach ridges that are significantly higher than surrounding ridges and are ideally paired with a tsunami sand sheet immediately behind them. 6. Conclusions We quantified morphological changes to a beach ridge plain in West Aceh in response to coseismic subsidence and postseismic uplift during and after the December 2004 Sumatra-Andaman earthquake and ensuing tsunami. Our observations show that beach ridge morphology is modified by large scale events and that a wider and more importantly, higher beach ridge forms in response to coseismic land level changes. The outsized ridge can be preserved within the stratigraphic record of beach ridge plains, if sediment supply is sufficient to counteract coastal profile adjustments due to rapid postseismic uplift in the years to decades following the event. We suggest that beach ridge morphology along active margins is useful to analyse episodic land level changes related to large megathrust earthquakes. In addition, the morphology of beach ridge plains can give clues to where paleotsunami deposits can be found. Tsunami deposits are usually thickest and best preserved in low-lying areas in the vicinity of the original shoreline. The identification of outsized beach ridges of seismic origin could help to recognize such pre-historic shorelines behind which tsunami deposits are most likely preserved. Author contributions KM, CKT, WF, BGM, EM, RA, NH, FJK, N, IR, and SUS did the fieldwork. BH, NH, and KM prepared the levelling data. CKT and KM conducted the shoreline change analysis. KM, CKT, BGM, SL, JEAS and DJW prepared the manuscript. Acknowledgements We would like to thank Jeff Freymueller for providing processed GPS data; Peter Ruggiero and Widjo Kongko for sharing bathymetric data, Carolin Ferwerda for helping with GIS and the Aceh and Nias Rehabilitation and Reconstruction Board (BRR) for providing aerial photographs of the study area. We greatly

11

appreciated the constructive reviews of Richard Briggs and an anonymous reviewer that largely improved this manuscript. We acknowledge funding from the Asian Studies Center at the University of Pittsburgh and the Mentorship Fund at the University of Pittsburgh at Johnstown. Wellesley College provided funding through the Faculty Awards Program and the Brachman Hoffman Small Grants Program. This study benefitted from a VIDI grant awarded to J. E. A. Storms through the Dutch Organization for Scientific Research (ALWeNWO, VIDI grant number 864.09.004) and was supported by the Deltares Research Program “Coastal, Estuarine and River Morphodynamics”.

References Anthony, E.J., 1995. Beach ridge development and sediment supply: examples from West Africa. Mar. Geol. 129, 175e186. Apotsos, A., Gelfenbaum, G., Jaffe, B., 2011. Process-based modeling of tsunami inundation and sediment transport. J. Geophys. Res. 116, F01006. http:// dx.doi.org/10.1029/2010JF001797. Atwater, B.F., 1987. Evidence for great Holocene earthquakes along the outer coast of Washington State. Science 236, 942e944. Atwater, B.F., Cisternas, M., Yulianto, E., Prendergast, A.L., Jankaew, K., Eipert, A.A., Fernando, W.I.S., Tejakusuma, I., Schiappacasse, I., Sawai, Y., 2013. The 1960 tsunami on beach-ridge plains near Maullín, Chile: landward descent, renewed breaches, aggraded fans, multiple predecessors. Andean Geol. 40 (3), 393e418. Ballarini, M., Wallinga, J., Murray, A.S., Oost, A.P., Bos, A.J.J., vanEik, C.W.E., 2003. Optical dating of young coastal dunes on a decadal scale time. Quat. Sci. Rev. 22, 1011e1017. Boak, E.H., Turner, I.L., 2005. Shoreline definition and detection: a review. J. Coast. Res. 21 (4), 688e703. Bookhagen, B., Echtler, H.P., Melnick, D., Strecker, M.R., Spencer, J.Q.G., 2006. Using uplifted Holocene beach berms for paleoseismic analysis on the Santa María Island, south-central Chile. Geophys. Res. Lett. 33, L15302. Briggs, R.W., Sieh, K., Amidon, W.H., Galetzka, J., Prayudi, D., Suprihanto, I., Sastra, N., Suwargadi, B., Natawidjaja, D., Farr, T.G., 2008. Persistent elastic behavior above a megathrust rupture patch: Nias island, West Sumatra. J. Geophys. Res. 113, B12406. Bristow, C.S., Pucillo, K., 2006. Quantifying rates of coastal progradation from sediment volume using GPR and OSL: the Holocene fill of Guichen Bay, southeast South Australia. Sedimentology 53, 769e788. Brooke, B., Ryan, D., Pietsch, T., Olley, J., Douglas, G., Packett, R., Radke, L., Flood, P., 2008. Influence of climate fluctuations and changes in catchment land use on Late Holocene and modern beach-ridge sedimentation on a tropical macrotidal coast: Keppel Bay, Queensland, Australia. Mar. Geol. 251, 195e208. Cameron, N., 1983. The Geology of the Takengon Quadrangle, Sumatra. In: Explanatory Note and Geological Map 1:250'000. Chlieh, M., Avouac, J.P., Hjorleifsdottir, V., Song, T.R.A., Ji, C., Sieh, K., Sladen, A., Hebert, H., Prawirodirdjo, L., Bock, Y., Galetzka, J., 2007. Coseismic slip and afterslip of the great Mw 9.15 Sumatra-Andaman earthquake of 2004. Bull. Seismol. Soc. Am. 97 (1A), S152eS173. Choowong, S., Phantuwongraj, T., Charoentitirar, V., Chutakositkanon, S., Yumuand, P., Charusiri, 2009. Beach recovery after 2004 Indian Ocean tsunami from Phang-nga, Thailand. Geomorphology 104, 134e142. n, F., Sawai, Y., Machuca, G., Lagos, M., Eipert, A., Cisternas, M., Atwater, B.F., Torrejo Youlton, C., Salgado, I., Kamataki, T., Shishikura, M., Rajendran, C.P., Malik, J.K., Rizal, Y., Husni, M., 2005. Predecessors of the giant 1960 Chile earthquake. Nature 437, 404e407. Engels, S., Roberts, M.C., 2005. The architecture of prograding sandy-gravel beach ridges formed during the last Holocene highstand: southwestern British Columbia, Canada. J. Sediment. Res. 75, 1052e1064. Finger, W., 2007. Hydrogeology of the Johan Pahlawan Subdistrict - Interpretation of the Geophysical Survey (2005) for the Assessment of Deep Ground Water Reserves. Final Report. Caritas, Switzerland, p. 27. Fritz, H.M., Blount, C., Sokoloski, R., Singleton, J., Fuggle, A., McAdoo, B.G., Moore, A.L., Grass, C., Tate, B., 2007. Hurricane Katrina storm surge distribution and field observations on the Mississippi Barrier Islands. Estuar. Coast. Shelf Sci. 74, 12e20. Graaff, R. de, 2007. Initial Modelling of Hydraulic Conditions. Aceh and Nias Sea Defense. Flood Protection, Refuges and Early Warning Project, BRR Concept Note/INFRA 300GI: WL j Delft Hydraulics Report. Goff, J., Sugawara, D., 2014. Seismic-driving of sand beach ridge formation in northern Honshu, Japan? Mar. Geol. http://dx.doi.org/10.1016/j.margeo.2014.04.005. Goy, J.L., Zazo, C., Dabrio, C.J., 2003. A beach-ridge progradation complex reflecting periodical sea-level and climate variability during the Holocene (Gulf of Almeria, Western Mediterranean). Geomorphology 50, 251e268. Hamzah, L., Puspito, N., Imamura, F., 2000. Tsunami catalogue and zones in Indonesia. J. Nat. Disaster Sci. 22, 25e43. Helland-Hansen, W., Hampson, G.J., 2009. Trajectory analysis: concepts and applications. Basin Res. 21, 454e483.

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014

12

K. Monecke et al. / Quaternary Science Reviews xxx (2014) 1e12

Holman, R.A., 1986. Extreme value statistics for wave run-up on a natural beach. Coast. Eng. 9, 527e544. Jaffe, B.E., Dengler, L., Hidayat, R., Kingsley, E., Kongko, W., Lukijanto, Moore, A., Titov, V., Yulianto, E., Borrero, J.C., Prasetya, G.S., Peters, R., McAdoo, B., Gelfenbaum, G., Morton, R., Ruggiero, P., Higman, B., 2006. Northwest Sumatra and offshore islands field survey after the December 2004 Indian Ocean tsunami. Earthq. spectra 22 (s3), 105e135. Jankaew, K., Atwater, B.F., Sawai, Y., Choowong, M., Charoentitirat, T., Martin, M., 2008. Medieval forewarning of the 2004 Indian Ocean tsunami in Thailand. Nature 455 (7217), 1228e1231. Liew, S.C., Gupta, A., Wong, P.P., Kwoh, L.L., 2010. Recovery from a large tsunami mapped over time: the Aceh Coast, Sumatra. Geomorphology 114 (4), 520e529. Masselink, G., Hughes, M.G., Knight, J., 2011. Introduction to Coastal Processes and Geomorphology, second ed. Hodder Education, London, p. 416. McAdoo, B.G., Dengler, L., Prasetya, G., Titov, V., 2006. Smong: How an oral history saved thousands on Indonesia's simeulue island during the december 2004 and march 2005 tsunamis. Earthq. Spectra 22 (S3), 661e669. Meilianda, E., Dohmen-Janssen, C., Maathuis, B., Hulscher, S., Mulder, J., 2010. Shortterm morphological responses and developments of Banda Aceh coast, Sumatra Island, Indonesia after the tsunami on 26 December 2004. Mar. Geol. 275 (1), 96e109. Meltzner, A.J., Sieh, K., Abrams, M., Agnew, D.C., Hudnut, K.W., Avouac, J.-P., Natawidjaja, D.H., 2006. Uplift and subsidence associated with the great AcehAndaman earthquake of 2004. J. Geophys. Res. 111, B02407. Meltzner, A.J., Sieh, K., Chiang, H.-W., Shen, C.-C., Suwargadi, B.W., Natawidjaja, D.H., Philibosian, B.E., Briggs, R.W., Galetzka, J., 2010. Coral evidence for earthquake recurrence and an A.D. 1390-1455 cluster at the south end of the 2004 Aceh e Andaman rupture. J. Geophys. Res. 115, B10402. Monecke, K., Finger, W., Klarer, D., Kongko, W., McAdoo, B.G., Moore, A.L., Sudrajat, S.U., 2008. A 1,000-year sediment record of tsunami recurrence in northern Sumatra. Nature 455 (7217), 1232e1234. Moore, A., Nishimura, Y., Gelfenbaum, G., Kamataki, T., Triyono, R., 2006. Sedimentary deposits of the 26 December 2004 tsunami on the northwest coast of Aceh, Indonesia. Earth Planets Space 58, 253e258. Moore, L.J., Ruggiero, P., List, J.H., 2006. Comparing high water and high water line shorelines: should proxy-datum offsets be incorporated into shoreline change analysis? J. Coast. Res. 22 (4), 894e905. Moore, L.J., 2000. Shoreline mapping techniques. J. Coast. Res. 16, 111e124. Nanayama, F., Satake, K., Furukawa, R., Shimokawa, K., Shigeno, K., Atwater, B.F., 2003. Unusually large earthquakes inferred from tsunami deposits along the Kuril Trench. Nature 424, 660e663. Natawidjaja, D.H., Sieh, K., Galetzka, J., Suwargadi, B.W., Cheng, H., Edwards, R.L., Chlieh, M., 2007. Interseismic deformation above the Sunda Megathrust recorded in coral microatolls of the Mentawai islands, West Sumatra. J. Geophys. Res. 112, B02404. Newcomb, K.R., McCann, W.R., 1987. Seismic history and seismotectonics of the Sunda Arc. J. Geophys. Res. 92, 421e439. Nurjanah, 2013. Aceh Paleotsunami Reconstruction for Disaster Risk Reduction and Global Information (Unpublished master thesis). Tsunami Disaster Mitigation and Research Center (TDMRC), Banda Aceh, Indonesia. Otvos, E.G., 2000. Beach ridges definitions and significance. Geomorphology 32, 83e108. Pajak, M.J., Leatherman, S., 2002. The highwater line as shoreline indicator. J. Coast. Res. 18 (2), 329e337. Pinegina, T.K., Bourgeois, J., 2001. Historical and paleo-tsunami deposits on Kamchatka, Russia: long-term chronologies and long-distance correlations. Nat. Hazards Earth Syst. Sci. 1, 177e185.

Plafker, G., 1972. Alaskan earthquake of 1964 and Chilean earthquake of 1960: implications for arc tectonics. J. Geophys. Res. 77, 901e925. Ruggiero, P., Komar, P.D., McDougal, W.G., Marra, J.J., Beach, R.A., 2001. Wave runup, extreme water levels and the erosion of properties backing beaches. J. Coast. Res. 17 (2), 407e419. Savage, J.C., Plafker, G., 1991. Tide gage measurements of uplift along the South Coast of Alaska. J. Geophys. Res. 96 (B3), 4325e4335. €ne, T., Illigner, J., Manurung, P., Subarya, C., Khafid, Zech, C., Galas, R., 2011. GPSScho controlled tide gauges in Indonesia e a German contribution to Indonesia's tsunami early warning System. Nat. Hazards Earth Syst. Sci. 11, 731e740. Shafer Rogers, S., Sandweiss, D.H., Maasch, K.A., Belknap, D.F., Agouris, P., 2004. Coastal change and beach ridges along the northwest coast of Peru: Image and GIS analysis of the Chira, Piura, and Cola'n Beach-Ridge Plains. J. Coast. Res. 20, 1102e1125. Shearer, P., Bürgmann, R., 2010. Lessons learned from the 2004 Sumatra-Andaman megathrust rupture. Annu. Rev. Earth Planet. Sci. 38 (1), 103e131. Shennan, Ian, Bruhn, R., Barlow, N.L.M., Good, K., Hocking, E., 2014. Late Holocene great earthquakes in the eastern part of the Aleutian megathrust. Quat. Sci. Rev. 84, 86e97. €ller, I., Evans, B.R., 2014. Where local Matters: impacts Spencer, T., Brooks, S.M., Mo of a major North sea storm surge. Eos Trans. AGU 95 (30), 269. Storms, J.E.A., Kroonenberg, S.B., 2007. The impact of rapid sea level changes on recent Azerbaijan Beach ridges. J. Coast. Res. 23 (2), 521e527. Subarya, C., Chlieh, M., Prawirodirdjo, L., Avouac, J.P., Bock, Y., Sieh, K., Meltzner, A.J., Natawidjaja, D.H., McCaffrey, R., 2006. Plate-boundary deformation associated with the great Sumatra-Andaman earthquake. Nature 440 (7080), 46e51. Suito, H., Freymueller, J.T., 2009. A viscoelastic and afterslip postseismic deformation model for the 1964 Alaska earthquake. J. Geophys. Res. 114, 404e426. Tamura, 2012. Beach ridges and prograded beach deposits as paleoenvironment records. Earth Sci. Rev. 114, 279e297. Tappin, D.R., Evans, H.M., Jordan, C.J., Richmond, B., Sugawara, D., Goto, K., 2012. Coastal changes in the Sendai area from the impact of the 2011 Tohoku-oki tsunami: interpretations of time series satellite images, helicopter-borne video footage and field observations. Sediment. Geol. 282, 151e174. Tanner, W.F., 1995. Origin of beach ridges and swales. Mar. Geol. 129, 149e161. Taylor, M., Stone, G.W., 1996. Beach-ridges: a review. J. Coast. Res. 12 (3), 612e621. Thieler, E.R., Himmelstoss, E.A., Zichichi, J.L., Ergul, A., 2009. Digital Shoreline Analysis System (DSAS) Version 4.0d an ArcGIS Extension for Calculating Shoreline Change. U.S. Geological Survey Open-File Report 2008e1278. Thompson, T.A., 1992. Beach-ridge development and lake-level variation in southern Lake Michigan. Sediment. Geol. 80, 305e318. Thompson, T.A., Baedke, S.J., 1995. Beach-ridge development in Lake Michigan: shoreline behavior in response to quasi-periodic lake-level events. Mar. Geol. 129, 163e174. Vargas, G., Farías, M., Carretier, S., Tassara, A., Baize, S., Melnick, D., 2011. Coastal uplift and tsunami effects asociated to the 2010 Mw 8.8 Maule earthquake in Central Chile. Andean Geol. 38 (1), 219e238. Wang, K., Hu, Y., He, J., 2012. Deformation cycles of subduction earthquakes in a viscoelastic earth. Nature 484, 327e332. Wang, P., Kirby, J.H., Haber, J.D., Horwitz, M.H., Knorr, P.O., Krock, J.R., 2006. Morphological and sedimentological impacts of Hurricane Ivan and immediate poststorm beach recovery along the Northwestern Florida Barrier-Island Coasts. J. Coast. Res. 22 (6), 1382e1402. Wells, A., Goff, J., 2007. Coastal dunes in Westland, New Zealand, provide a record of paleoseismic activity on the Alpine fault. Geology 35, 731e734. Yu, F., Switzer, A.D., Lau, A.Y.A., Yeung, H.Y.E., Chik, S.W., Chiu, H.C., Huang, Z., Pile, J., 2013. A comparison of the post-storm recovery of two sandy beaches on Hong Kong Island, southern China. Quat. Int. 304, 163e175.

Please cite this article in press as: Monecke, K., et al., Beach ridge patterns in West Aceh, Indonesia, and their response to large earthquakes along the northern Sunda trench, Quaternary Science Reviews (2014), http://dx.doi.org/10.1016/j.quascirev.2014.10.014