Quaternary Science Reviews 122 (2015) 258e281

Contents lists available at ScienceDirect

Quaternary Science Reviews

journal homepage: www.elsevier.com/locate/quascirev

Time-varying interseismic strain rates and similar seismic ruptures on the NiaseSimeulue patch of the

* Aron J. Meltzner a, b, , Kerry Sieh a, b, Hong-Wei Chiang a, c, Chung-Che Wu c, Louisa L.H. Tsang a, Chuan-Chou Shen c, Emma M. Hill a, Bambang W. Suwargadi d, Danny H. Natawidjaja d, Belle Philibosian b, e, Richard W. Briggs f a Earth Observatory of Singapore, Nanyang Technological University, 639798, Singapore b Tectonics Observatory, California Institute of Technology, Pasadena, CA 91125, USA c High-precision Mass Spectrometry and Environment Change Laboratory (HISPEC), Department of Geosciences, National Taiwan University, Taipei 10617, Taiwan, ROC d Research Center for Geotechnology, Indonesian Institute of Sciences (LIPI), Bandung 40135, Indonesia e Equipe de Tectonique et Mecanique de la Lithosphere, Institut de Physique du Globe de Paris, 75238 Paris, France f Geologic Hazards Science Center, U.S. Geological Survey, Denver, CO 80225, USA article info abstract

Article history: Fossil coral microatolls from fringing reefs above the great (MW 8.6) megathrust rupture of 2005 record Received 11 March 2015 uplift during the historically reported great earthquake of 1861. Such evidence spans nearly the entire Received in revised form 400-km strike length of the 2005 rupture, which was previously shown to be bounded by two persistent 22 May 2015 barriers to seismic rupture. Moreover, at sites where we have constrained the 1861 uplift amplitude, it is Accepted 4 June 2015 comparable to uplift in 2005. Thus the 1861 and 2005 ruptures appear to be similar in both extent and Available online xxx magnitude. At one site an uplift around AD 1422 also appears to mimic the amount of uplift in 2005. The high degree of similarity among certain ruptures of this NiaseSimeulue section of the Sunda megathrust Keywords: zone contrasts with the substantial disparities amongst ruptures along other sections of the Sumatran portion Paleoseismology of the Sunda megathrust. At a site on the northwestern tip of , reefs also rose during an earthquake Paleogeodesy in AD 1843, known historically for its damaging along the eastern coast of the island. The coral microatolls also record interseismic vertical deformation, at annual to decadal resolution, Coral microatolls spanning decades to more than a century before each earthquake. The corals demonstrate significant Earthquake recurrence changes over time in the rates of interseismic deformation. On southern , interseismic subsi- Interseismic deformation dence rates were low between 1740 and 1820 but abruptly increased by a factor of 4e10, two to four Slow slip events decades before the 1861 rupture. This may indicate that full coupling or deep locking of the megathrust began only a few decades before the great earthquake. In the Banyak Islands, near the pivot line sepa- rating coseismic uplift from subsidence in 2005, ongoing interseismic subsidence switched to steady uplift from 1966 until 1981, suggesting a 15-year-long slow slip event, with slip velocities at more than 120% of the plate convergence rate. © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction earthquake recurrence in a region, the more that region can pre- pare for the hazards it faces. And the more complete we can make Assessing future earthquake hazard relies upon an appreciation our picture of strain accumulation, and how strain accumulation for the range of earthquake scenarios that are plausible for a varies over time, the better our chances for accurately identifying particular and an understanding of the strain accumulation faults that are likely to rupture in the near future. history along that fault. The better we can characterize the There have been limited efforts to apply earthquake recurrence models to subduction megathrusts. Few long paleoseismic records exist for subduction zones with which to rigorously test these * Corresponding author. Earth Observatory of Singapore, Nanyang Technological models, and the inaccessibility of megathrusts hinders attempts to University, 639798, Singapore. compare displacements at a point along the fault from one event to E-mail address: [email protected] (A.J. Meltzner). http://dx.doi.org/10.1016/j.quascirev.2015.06.003 0277-3791/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 259 the next. In Sumatra, prior studies identified two persistent barriers number of settings (Beroza and Ide, 2009; Gomberg and The to rupture, under the Batu Islands (Natawidjaja et al., 2006) and Cascadia 2007 and Beyond Working Group, 2010; Peng and under central Simeulue (Meltzner et al., 2012)(Fig. 1). These two Gomberg, 2010). Multiple large SSEs, with durations of 2e4 years, barriers, which align with fracture zones in the subducting slab, and a series of abrupt changes in the width of the locked region, divide the Sumatran portion of the Sunda megathrust into at least have now been documented in southern Alaska (Fu and three segments with independent rupture histories. North of the Freymueller, 2013; Freymueller et al., 2014; Fu et al., 2014). In the northern barrier (on the Aceh segment) and south of the southern Tokai region of Japan, a 5-year-long SSE occurred between 2000 barrier (on the Mentawai segment), paleoseismic evidence sug- and 2005, and longer-term changes in plate coupling have been gests that ruptures vary considerably: no two ruptures in the observed (Ochi and Kato, 2013). Changes in plate coupling over available paleoseismic, historical, or modern records even vaguely time have also been proposed elsewhere (Nishimura et al., 2004; resemble one another (Meltzner et al., 2010; Philibosian et al., Prawirodirdjo et al., 2010; Ozawa et al., 2012; Uchida and 2014). The 28 March 2005 MW 8.6 rupture spanned the full dis- Matsuzawa, 2013; Mavrommatis et al., 2014; Philibosian et al., tance between these two barriers, and since both rupture end- 2014; Yokota and Koketsu, 2015). points appear to have been structurally controlled, we speculate What if a fault system can appear for decades to be uncoupled that earthquakes like 2005 may be a common feature of this and then suddenly start accumulating strain that could lead to portion of the megathrust. seismic rupture? If this could happen, it would have profound As for fault behavior between earthquakes, researchers gener- implications for hazards along subduction zones and other faults ally believed until recently that interseismic motions are roughly that are not currently considered highly seismogenic. Interseismic linear over time, punctuated only by sudden earthquakes and deformation rates, long assumed to be steady over time, may postseismic deformation that follows the earthquakes (Savage and instead be a function of time. Most modern geodetic networks have Thatcher, 1992). Although postseismic transients in deformation not been in operation for sufficiently long durations to address this have been widely documented (e.g., Melbourne et al., 2002; Zweck question. The geological record may provide unique insight. et al., 2002; Hu et al., 2004; Sawai et al., 2004; Hsu et al., 2006; In this paper, we explore the recent paleoseismic (earthquake) Pollitz et al., 2008; Perfettini et al., 2010; Hu and Wang, 2012) histories of sites on Nias, Bangkaru, and southern (eastern) and result from a variety of processes during the post-earthquake Simeulue islands, which lie above the 28 March 2005 MW 8.6 deformation phase of the earthquake cycle (Perfettini et al., 2005; rupture patch (Briggs et al., 2006)(Figs. 1e2). We combine histor- Wang et al., 2012; Bürgmann and Thatcher, 2013; Sun et al., ical records with geological observations from in situ preserved 2014), they are commonly observed to decay, over a period of coral coloniesdnamely coral microatollsdto determine details of years to decades, to a “background” interseismic rate. The belief the timing, extent, and magnitude of past coseismic deformation. was that, subsequently, this “background” interseismic strain rate These data elucidate similarities and differences between various (or pattern of interseismic deformation) remained steady over past earthquakes, including notable similarities between earth- most of the seismic cycle (Savage and Thatcher, 1992). More quakes in 1861 and 2005. We also explore the recent paleogeodetic recently, researchers discovered processes and phenomena previ- (interseismic deformation) histories of these sites. The coral ously unappreciated along subduction zones. Numerous studies microatolls provide information on gradual relative sea-level (RSL) have explored slow slip events (SSEs) at a range of timescales, in a change (hence land-level change) between earthquakes, which we

Fig. 1. (a) Regional map of the Sunda megathrust and large megathrust ruptures since AD 1865. Rupture locations and magnitudes are from Briggs et al. (2006), Konca et al. (2008), Meltzner et al. (2010), Hill et al. (2012), and references therein; the 1907 location is speculative. Relative plate motions from Shearer and Bürgmann (2010). Black lines are faults; gray lines are fracture zones. (b) The 2005 rupture spanned the NiaseSouthern Simeulue segment of the megathrust (NSS), which is bounded by persistent rupture barriers (gray bars). Ruptures to the north on the Aceh segment (A) and to the south on the Mentawai segment (M) have been highly variable, but there may be a higher degree of similarity among the largest ruptures along the NiaseSouthern Simeulue segment. Bdg, Badgugu; Smk, Simuk Island. 260 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 2. Map of sites with coral data published in this and earlier papers. Of the sites over the 2005 rupture patch, some show evidence for deformation in 1843 and others have evidence for uplift in 1861. No site has direct evidence for uplift in both earthquakes, although sites with 1843 deformation are also inferred to have risen in 1861. Data from sites to the northwest, over the 2004 patch and the boundary region between the two ruptures, were published in earlier papers. Contours show uplift and subsidence in the 2005 earthquake, updated from Briggs et al. (2006) and Meltzner et al. (2012). Sites published by Meltzner et al. (2010): USL (Ujung Salang), LDL (Lhok Dalam), LNG (Langi), LKP (Lhok Pauh), LWK (Lewak), USG (Ujung Sanggiran), PST (Pulau Salaut Besar). Sites published by Meltzner et al. (2012): ULB (Ujung Lambajo), BUN (Bunon), PPY (Pulau Penyu). Sites with microatolls analyzed in this paper: SLR (Silinggar), SMB (Sambay), UTG (Ujung Tinggi), LBJ (Labuhan Bajau), LAT (Latiung), PBK (Pulau Bangkaru), PWG (Pulau Wunga), AFL (Afulu), PSN (Pulau Senau), MZL (Muzoi Ilir), BWL (Bawelowalani), LAG (Lagundri). Additional sites with data listed in the supplementary tables of this paper: SBG (Sinabang), GSG (Teluk Gosong, or Busung), SBA (Siaba). can use to infer rates of interseismic deformation and patterns of along the west coast of Nias, does not necessarily indicate that strain accumulation. These corals reveal that rates of interseismic other areas were unaffected. Even today, the island's west coast is vertical deformation are not constant over time. rugged and sparsely inhabited, and a large tsunami there in 1843 would not necessarily have left a historical record. Prior to this 2. Historical accounts of earthquakes since 1843 study, no land-level changes were attributed to the 1843 earthquake. Limited historical information is available for three large In February 1861, a strong and widely felt earthquake affected earthquakes in the NiaseSimeulue region prior to 2005 (see Ap- Nias and northern Sumatra. Tsunami inundation was reported pendix). The earliest historical event, in January 1843, caused se- along the southwest and east coasts of Nias, in the Batu Islands, and vere shaking on Nias and a substantial tsunami that inundated (at in numerous places along the coast of mainland northern Sumatra minimum) the northeast coast of Nias and reached the adjacent (The Singapore Free Press, 1861a, 1861b; Zurcher and Margolle, mainland coast (Baird Smith, 1845; Junghuhn, 1845; Wichmann, 1866, 1868; Wichmann, 1922). Reports from the time unequivo- 1918). Few additional details are known historically about this cally describe coseismic uplift of some parts of the west coast of earthquake; the lack of more widely reported effects, particularly Nias and permanent flooding (subsidence, slumping, or sediment A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 261 compaction) along other portions of the west coast of Nias and in variability of the amplitude of the diedown in coeval corals at Singkil, on the adjacent mainland coast (Fig. 2)(The Singapore Free nearby sites (the amplitude of tectonic uplifts tends to vary mark- Press, 1861b; Zurcher and Margolle, 1866, 1868; Wichmann, 1922). edly over short distances, whereas IOD-related diedowns should be In January 1907, a , with an estimated similar over distances of tens to hundreds of kilometers) and the magnitude of MS 7.5 to 8.0, appears to have involved rupture of the duration of the RSL change (the IOD causes fluctuations in RSL shallow, updip portion of the NiaseSouthern Simeulue segment of lasting weeks to months, whereas tectonic changes are more the megathrust (Kanamori et al., 2010). This event produced strong enduring, lasting decades to centuries). Meltzner et al. (2010) shaking on Simeulue and Nias and a tsunami that devastated suggested that a moderate non-tectonic diedown on a coral Simeulue, Nias, and the Batu Islands and extended 950 km along should be followed by unrestricted upward growth without addi- the mainland Sumatra coast (Newcomb and McCann, 1987). News tional diedowns until the coral grows back up to its former eleva- accounts, although prone to exaggeration, stated that the southern tion. Meltzner and Woodroffe (2015) provide further discussion of coast of Simeulue was destroyed, and that the island “nearly dis- techniques to differentiate uplift from transient oceanographically appeared” underwater (The Sydney Morning Herald, 1907). The induced diedowns. Malacca Strait Pilot (Great Britain Hydrographic Department, 1934), a nautical guidebook concerned with navigation, anchorage, and 3.2. Microatoll records of gradual RSL change (paleogeodesy) bathymetric depths, indicates that the southern coast of Simeulue “was partially submerged by an earthquake” in 1907. A microatoll's basic morphology reveals important information The March 2005 earthquake involved uplift of southern about RSL during the coral's lifetime. Flat-topped microatolls record (eastern) Simeulue, Bangkaru, and most of Nias, and subsidence of RSL stability; colonies with diedowns (HLS unconformities) that the eastern Banyak Islands and easternmost Nias (Briggs et al., rise radially outward toward their perimeter reflect rising sea level 2006). Uplift in 2005 peaked at 290 cm at Lahewa on the north- during their decades of growth. As reefs subside or rise in the western tip of Nias; uplift exceeding 50 cm extended from the course of tectonic elastic strain accumulation and release, micro- southernmost west coast of Nias to the island's northern tip and atoll morphologies record changes in RSL. Because these corals' northward to the eastern half of Simeulue; and the southwestern skeletons have annual growth bands, we can precisely calculate half of Bangkaru also uplifted >50 cm (Fig. 2). At Simuk Island, just rates of change in elevation, when those changes are gradual. south of the Equator in the Batu Islands (Fig. 1), 25 cm of uplift was In order to determine gradual (interseismic) land-level changes, recorded (Briggs et al., 2006). The endpoints of the 400-km-long we first estimate rates of RSL change (and associated errors) from 2005 rupture coincide with the persistent rupture barriers of the coral growth histories, following Meltzner et al. (2010, 2012). Meltzner et al. (2012) (Fig. 1). Unless a compelling argument can be made otherwise, a “worst- case scenario” is considered in which there is an 8-cm error in the 3. Coral microatoll background and methodology apparent elevation gain recorded by a microatoll slab due to dif- ferential erosion of one part of the coral compared to another part, 3.1. Limits on coral upward growth: diedowns or due to deficient upward growth. The error in the rate of RSL change, then, is 8 cm divided by the length of the record. We extracted records of RSL change from coral microatolls. In If sea level itself was steady as the coral grew, then the land- the absence of reef ponding (which in general has not been level change is simply the opposite of RSL change. Alternatively, if observed on the typically narrow reef flats of Simeulue or Nias), the rate of sea-level rise or fall was not negligible but is known coral microatolls grow upward to a limit near mean low water (within error), it can be subtracted from the overall rate of RSL springs (MLWS), and their upper surfaces record a history of RSL change before the negative of that rate is used to calculate land- (Scoffin and Stoddart, 1978; Taylor et al., 1987; Zachariasen et al., level change. To the extent that rates of regional sea-level change 2000; Meltzner et al., 2010). Microatoll shapes form because pro- are unknown at various times in the past, this adds uncertainty to longed subaerial exposure at times of extreme low water limits the our estimates of land-level change. However, we also consider the highest level to which the coral colonies can grow (Briggs et al., spatial scales over which SSH trends vary: analyses of satellite 2006; Meltzner et al., 2010; Meltzner and Woodroffe, 2015) altimetry data since 1993 suggest that sea-level trends vary fairly (Figs. 3e4). A diedown to a uniform elevation around the perimeter smoothly at low latitudes. In particular, average sea-level rise of the coral is a clear indication that the diedown resulted from low calculated over the period 1993e2009 varied only from ~2.0 mm/yr water, and the elevation above which all coral died is termed the just northwest of Simeulue, to ~2.5 mm/yr just southeast of highest level of survival (HLS) (Taylor et al., 1987). A related term, Simeulue, to ~3.0 mm/yr near Nias and the Batu Islands, a distance the highest level of growth (HLG), reflects the highest elevation up of 600 km (Beckley et al., 2007; Hamlington et al., 2011). If this to which a coral grew in a given year. Although both HLS and HLG spatial variability in SSH trends since 1993 is characteristic of the refer to the highest living coral at a particular time of interest, HLG spatial variability in SSH trends in the same region at earlier times, is limited by a coral's upward growth rate. Hence, in years during it puts a limit on how much variability among coeval RSL records which there is no diedown, HLG provides only a minimum estimate from corals at nearby sites can be explained by spatial variability in of the HLS that would theoretically be possible, given water levels. SSH trends. In other words, if an abrupt and sustained change in RSL Any coral diedown at sites off the west coast of Sumatra may be trends of >1 mm/yr occurs at one site but not contemporaneously related to tectonic uplift, the Dipole (IOD), or both. at another site 10e100 km away, it is unlikely that this is the result Positive IOD events result in the development of persistent surface of changes in SSH trends. easterly winds over the equatorial Indian Ocean, and lower sea surface height (SSH) in the tropical eastern Indian Ocean 3.3. Coseismic uplift inferred from sudden RSL fall (Chambers et al.,1999; Webster et al.,1999; Yu and Rienecker, 1999; (paleoseismology) Brown et al., 2002; van Woesik, 2004; Meltzner et al., 2010), whereas negative IOD events have the opposite effect. If a diedown For sudden uplifts inferred from diedowns, we attempt to esti- is sufficiently large, it is unlikely to result solely from IOD effects mate formal errors. Aside from any uncertainty in the amplitude of and is more likely to be related to tectonics. Other criteria for dis- the diedown that may result from erosion of the microatoll, there tinguishing uplift from IOD-related diedowns include the spatial are two primary sources of uncertainty in estimating the uplift. The 262 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

t = year 8 t = year 11 t = year 11.1 t = year 14 t = year 14.1 t = year 17

Case 1

Case 2

Case 3

t = year 17 t = year 17.1 t = year 20 t = year 20.1 t = year 23

Case 4 annual banding sea level Sudden LEGEND Tectonic living perimeter short-lived period of lower sea level Uplift

Sudden Tectonic Case 5 Subsidence

Fig. 3. (Top) Microatoll development under different relative sea-level history scenarios: under stable sea level conditions (Case 1), under gradually falling sea level (Case 2), and under gradually rising sea level (Case 3). In all three cases, we superimpose on the long-term trend a realistic interannual variability: at 11.1 and 14.1 yr, we simulate temporary local sea level lowerings as might occur during positive Indian Ocean Dipole (IOD) events. Concentric annuli form simply from year-to-year fluctuations in low water level. (Bottom) Microatoll development affected by sudden changes in land level. Case 4 illustrates the microatoll from Case 3, followed by coseismic uplift at 17.1 yr; Case 5 illustrates the microatoll from Case 1, followed by coseismic subsidence at 17.1 yr. In each case, the long-term trend is superimposed on a typical IOD cycle, with an additional period of lower local sea level at 20.1 yr. In Case 4, the uplift must have been sudden at 17.1 yr, but if we found the microatoll in Case 5, we could not distinguish between sudden coseismic subsidence at 17.1 yr and rapid interseismic subsidence (at an average rate exceeding the coral's growth rate) beginning at 17.1 yr.

Fig. 4. Photos of microatolls that died in 1861 on southeastern Simeulue. Both exhibit a wide, low-relief interior that is surrounded by concentric “stair step” rings rising outward. (a) LBJ-2, Labuhan Bajau site, diameter ~5.0 m. (b) LAT-1, Latiung site, diameter ~3.5 m.

first is the inherent variability in the corals' HLG or HLS. In the The amplitude of the diedown is generally treated as a proxy for Mentawai Islands, the variation of HLS on a single Porites microatoll the amount of uplift, but there is an additional source of uncer- is usually about ±2.6 cm (2s)(Zachariasen et al., 2000; Natawidjaja tainty that is incurred in this conversion: inherent variability in SSH et al., 2004), which is consistent with our observations farther associated with phenomena such as the IOD (Fig. 3). Diedowns north on Simeulue and Nias. Hence, for diedowns measured on a unrelated to tectonic uplift during the 1961 and 1997 positive IOD single microatoll, where neither the pre-diedown HLG nor the eventsdtwo of the strongest on recorddreached 10 cm and 12 cm, post-diedown HLS are significantly eroded, the uncertainty should respectively, at sites on Simeulue, whereas diedowns unrelated to be roughly [(2.6 cm)2 þ (2.6 cm)2]1/2, or less than about ±4 cm. uplift during the more moderate positive IOD events in 1982 and A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 263

1991 reached 9 cm and 5 cm at sites on Simeulue (Meltzner et al., obtained uplift data for their joint inversion of slip in the 2005 2010). These observations suggest minimum uncertainties of rupture. All of these sites rose >60 cm in the 2005 earthquake, ±12 cm in converting diedowns to uplift at sites off the west coast except for SLR (Silinggar), near the northern limit of 2005 uplift; of Sumatra, if nothing is known about the SSH history at the loca- LAG (Lagundri), near the southeastern limit of uplift; and PBK tion. (These uncertainties may be larger in the Mentawai Islands (Pulau Bangkaru), near the eastern limit of uplift. south of the Equator, where IOD-related SSH variability tends to be There is direct evidence that the 1861 rupture did not extend larger.) farther south than the rupture in 2005. Immediately south of the That said, a record of the strength of the IOD since AD 1846 2005 rupture patch (where continuous GPS observations suggest (Abram et al., 2008) is helpful for the 1861 (and subsequent) there was little vertical change in 2005 (Briggs et al., 2006)), a long- earthquakes. The period from 1858 to 1871 was fairly quiet, without lived microatoll at the Badgugu (BDG) site in the Batu Islands any severe IOD events. This observation implies that conversions (Fig. 1) recorded a diedown of ~2 cm in 1861 (Philibosian et al., from measurements of 1861 diedowns to estimates of 1861 uplifts 2014); in other words, there was little vertical change near Badg- incur at most 4 or 5 cm of uncertainty. Combining this uncertainty ugu in either 1861 or 2005. The only large uplift at Badgugu in the with the inherent variability in corals' HLS mentioned earlier, we 260 years preceding 2005 was ~70 cm, during a moderate (MW 7.7) assign errors of ±6cm(2s) to our estimates of 1861 uplift, in earthquake in 1935, south of the 2005 rupture (Natawidjaja et al., optimal (uneroded) cases. 2004, 2006). There is weaker evidence that the 1861 rupture did not extend 3.4. Dating techniques and precision farther north than the rupture in 2005. Although no corals that could give us information about land-level change in 1861 were We can date the times of past uplift or subsidence events using ever found on the reef flats of northwestern Simeulue (a problem UeTh techniques, which optimally enable determination of the age discussed by Meltzner et al. (2010, 2012)), the lack of strong shaking of a coral sample to plus or minus a few years (Shen et al., 2002, or damage reported from northern Aceh argues that the north- 2003, 2008, 2012). In cases where individual samples yield insuf- western limit of rupture in 1861 did not extend substantially ficiently precise ages, we can obtain increasingly precise estimates beyond central Simeulue. Furthermore, the existence of a persistent of the age of a microatoll by establishing a weighted average of barrier that arrested rupture under central Simeulue in numerous dates from multiple samples from a single slab. The number of other earthquakes over the past 1100 years (Meltzner et al., 2012) annual bands separating the various samples is also considered in (Fig. 1), supports the inference that rupture in 1861 stopped there. this calculation (Meltzner and Woodroffe, 2015). We can improve upon the dating precision even further by 5. RSL change from corals in 1861 and 1843 comparing the diedowns and the intervals between the diedowns on the various microatolls. The most prominent and ubiquitous 5.1. Coral diedowns and inferred uplift in 1861 diedowns (e.g., those in late 1961, late 1982, and late 1997 in 20th- century microatolls, separated by 21 and 15 years, respectively; Corals at nine of our sites recorded diedowns in 1861. After those in mid-1817, early 1833, and mid-1846 in the pre-1861 carefully assessing the age of each microatoll by considering UeTh microatolls, separated by 15.5 and 13.5 years, respectively) serve dates, historical information, and correlations between diedowns at as the coral analogue of geological marker beds, in that, despite different sites (see details in the supplementary text), we identified their presence in separate corals at separate sites, they are so at least one microatoll with an outer preserved band that formed in distinctive that there is no doubt as to their equivalent age. Because 1858, 1859, or 1860 at sites SLR (Silinggar), SMB (Sambay), UTG large tectonic diedowns can be tied directly to the 1843 or 1861 (Ujung Tinggi), LBJ (Labuhan Bajau), PBK (Pulau Bangkaru), MZL earthquakes in several corals, and because “marker bed” diedowns (Muzoi Ilir), and LAG (Lagundri) (Fig. 2). It is reasonable to infer that can be correlated to one another in most corals, the effective un- these microatolls died due to uplift in early 1861 (e.g., Fig. 5), certainty in the age of most of our slabs is simply the uncertainty in because it is often the case that the outermost one or two annual counting bands forward or backward from these diedowns. The bands are removed by erosion in the years subsequent to death and diedown correlations are discussed further in the supplement, in exposure (Meltzner et al., 2010, 2012; Philibosian et al., 2014). At Text S1.2, and then on a site by site basis in Text S2 through S13. site PWG (Pulau Wunga), where the two fossil corals were visibly heavily eroded, the outer preserved bands grew in 1852 and 1853, 3.5. Methods, scope, and content of this paper respectively; we also infer these corals to have died in 1861. Although we were unable to collect slabs at the LAT (Latiung) site, We slabbed, x-rayed, and analyzed the coral microatolls, UeTh dates from chiseled hand samples suggest one population of following the methods described by Meltzner et al. (2010, 2012) corals there died in 1861, and those microatolls were morphologi- and Meltzner and Woodroffe (2015). Interpreted coral cross sec- cally similar to corals that died in 1861 at the nearby LBJ site (Fig. 4, tions and time series are presented as supplementary material S1). (Figures S1eS86). In this paper, we focus on corals that are inferred We can estimate the amount of uplift in 1861 precisely at two to have died in AD 1843 or thereafter, as corals that died around AD sites where at least some corals survived the diedown, and we can 1800 or earlier have been sampled too sparsely to make solid in- estimate minimum bounds on the uplift at several other sites. At terpretations at present. Tables S1eS3 list details and UeTh dating site PBK on Bangkaru Island, near the hinge line separating uplift results from all Nias, Bangkaru, and eastern Simeulue corals from subsidence in 2005 (Fig. 2), the 1861 diedown was recorded collected thus far. by a coral (PBK-7) that was barely below HLS before the earthquake and was in sufficiently deep water to survive the diedown around 4. Distribution of sites its base. We estimate the 1861 uplift at this site to be the amplitude of the diedown on this coral, 30 cm, plus up to 5 cm to account for The sites presented in this paper sit above the 2005 the estimated erosion of the top of the central hemisphere (Fig. 6). NiaseSimeulue rupture patch (Fig. 1), along or just off the coasts of At site LAG on southwestern Nias, microatolls LAG-1 and LAG-4 Nias, Bangkaru, and the eastern half of Simeulue (Fig. 2). Many of provide the best estimates of HLG just before the 1861 diedown these sites are colocated with sites at which Briggs et al. (2006) (the pre-diedown HLG), as they are the highest and least eroded 264 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 5. Cross sections through parts of (a) slab UTG-5 and (b) slab UTG-6, both from site UTG. The UeTh dates suggest both died in the historical 1861 earthquake, but UTG-5 has been moderately eroded and is missing ~2.5 bands; UTG-6 sustained less erosion and is missing only ~0.5 band. Both corals experienced diedowns in early 1833 and ~13.5 years later in mid-1846 that are inferred to have resulted from regional sea-level lowerings related to the IOD. An earlier diedown on UTG-6 may correlate with a mid-1817 diedown on UTG-5, but ambiguities in band counting on UTG-6 (where the banding is not shown in the figure) preclude a definitive correlation. Both cross sections are plotted at the same scale; note the finer banding in UTG-6, a Goniastrea sp. coral, than in UTG-5, a Porites sp. coral. For the full slabs, see Figures S16 and S17.

corals from that time (Table S3; Figures S75, S79). A nearby coral, preclude more than a few centimeters of coseismic subsidence in LAG-3B, appears to have survived the 1861 uplift by tilting and 1843, unless such subsidence was countered by a nearly equal settling during shaking in 1861: it recorded a diedown in 1861, but amount of postseismic uplift within the first 3 years. by lowering its elevation relative to the substrate at the moment of The history at Afulu (site AFL), on the west coast of Nias, is the uplift, its lower half survived when every other coral we slab- different. Microatoll AFL-3 was an exceptionally well-preserved bed at this site died entirely (Table S3; Figure S78). Although the fossil coral, with its detailed morphology still intact, a sign of pre-1861 HLG on LAG-3B is no longer at its original elevation minimal erosion. In the slab, its outer annual band maintains a because of the tilting and settling, the post-diedown HLS is a reli- uniform thickness, a further indication of negligible erosion that is able indicator of RSL immediately after the uplift. We estimate the rarely seen in fossil corals (Fig. 9a). After diedown correlation with 1861 uplift at LAG as the difference between the pre-diedown HLG corals from other sites, its outer preserved band dates to 1842. AFL- on LAG-1 and LAG-4 and the post-diedown HLS on LAG-3B, which 4, a mushroom-shaped non-microatoll Goniastrea coral, was found is 28 cm (Fig. 7), and then add 6 cm to account for the estimated upright on its thin “stalk” in a position that made it unlikely to have erosion of the outer rims of LAG-1 and LAG-4; this yields an esti- ever been rolled or otherwise transported. Located 150 m seaward mate of ~34 cm. of AFL-3, our initial field interpretation was that AFL-4 was in situ At all other sites where corals died in 1861, we can place a and killed by the same uplift as AFL-3. AFL-4 was also minimally minimum bound on the amount of uplift that occurred in 1861 eroded, with a remarkably pristine outer surface and an outer based on the height of what was the living outer perimeter of the preserved band of nearly uniform thickness (Fig. 9b). A lack of coral at the time, coupled with the assumption that the coral was diedowns recorded by this coral (because it was lower in the water) killed entirely by uplift. In several cases, the minimum bound is not precludes diedown correlation to microatolls at this or nearby sites, very useful because the living perimeter of the coral was so short, but the two UeTh dates from AFL-4 are both remarkably precise while in all likelihood the uplift was large. Nonetheless, we provide and agree with one another, giving an estimate of the age of the estimates of 1861 uplift at various sites in Fig. 8 and in Table 1. A few coral's outer band as AD 1843.6 ± 2.0 (Table S3), an estimate of the more useful minimum estimates for 1861 uplift are 45 cm completely independent of that for the age of AFL-3. Together these at SLR; 110 cm at LBJ; and 180 cm at PWG (see details in sup- observations offer compelling evidence for uplift at AFL in 1843. plementary text). The height of what was then the living outer perimeter of AFL-3 is 17 cm (Fig. 9a); for AFL-3 to have died entirely as a result of uplift, 5.2. RSL change and inferred land-level change in 1843 that uplift must have been at least 17 cm. The history at Pulau Senau (site PSN), north and landward of site In contrast to the widespread uplifts in 1861, corals at most of AFL, is also different from all other sites. PSN-2 was more heavily our sites rule out significant land-level change in 1843. No fossil eroded than the AFL fossil corals, and its outer band, after diedown coral at site SLR, SMB, UTG, LBJ, PBK, PWG, MZL, or LAG experienced correlation with corals from other sites, dates to 1848 (Fig. 10). It is a diedown in 1842 or 1843 that could be attributed to the January conceivable, given the eroded condition of the outer surface of this 1843 earthquake, even allowing for a full 1-year band-counting microatoll, that 12 annual bands have been completely eroded, but uncertainty. Minor diedowns on many of the corals a few years the approximate radial symmetry of the coral and its pattern of later in mid-1846, inferred to be related to the IOD (e.g., Fig. 5), erosion suggest that a more likely explanation is that this coral died A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 265

Fig. 6. Fossil corals from the PBK site provide information not only on the timing of the predecessor to the 2005 earthquake, but also on the amount of uplift resulting from that earlier rupture. (a) Cross section through part of slab PBK-7, from subsite PBK-B. For the full slab, see Figure S37. (b) Relative sea-level history (coral growth history) for subsite PBK-B, derived from slabs PBK-5, PBK-7, and PBK-8. Different colors represent data from different corals. PBK-7 had almost reached HLS just prior to the 1861 earthquake (determined from its elevation relative to coeval microatolls at the site), and it survived the 1861 diedown. We estimate the 1861 uplift at this site to be the amplitude of the diedown on PBK-7, 30 cm, plus up to 5 cm to account for erosion of the top of the central hemisphere. Between 1751 and 1861, the corals record an average rate of RSL rise of 2.2 ± 0.7 mm/yr. That RSL rise appears to have been fairly steady from 1812 onward (2.5 ± 1.6 mm/yr if only data from 1812 to 1861 are considered), but there is ambiguity in the RSL history before 1812 because of erosion of a portion of PBK-5: we can neither preclude nor confirm decadal-scale fluctuations in the rate of RSL change prior to 1812. prior to 1861 for reasons not related to uplift. That enigma land-level changes in 1843: site AFL uplifted coseismically, whereas notwithstanding, two important observations regard diedowns site PSN, above a more downdip portion of the megathrust than the that did not happen on this microatoll. First, unlike at AFL, no portion under AFL, subsided either coseismically or postseismically diedown occurred at PSN around the time of the 1843 earthquake. within the first 1e3 years after the earthquake. An uplift farther Second, no IOD-related diedown occurred in 1846, either. PSN is the south at BWL may have occurred in the 1843 earthquake, but am- only site on Nias with a living coral in 1846 but with no suggestion biguity remains due to imprecision of the dates (Figure S73; of a diedown during that year. Despite considerable erosion of the Table S3). Whether the 1843 deformation was caused by rupture outer surface of PSN-2, the highest part of the 1846 annual band is along the megathrust or a splay fault is ambiguous, as is the preserved, so it is not possible for evidence of such a diedown to mechanism relating the deformation to the reported tsunami. We have been destroyed by erosion. One likely explanation for the cannot say based on observations at AFL or PSN alone whether any missing 1846 diedown at site PSN would be if there had been more uplift occurred at those sites in 1861, but based on 1861 uplift both subsidence at PSN in the preceding years than at other sites in our trenchward at PWG and landward at MZL (Fig. 2), it is very likely study area. We propose that such subsidence did occur at PSN, that both AFL and PSN rose in 1861. either coseismically in 1843, or as a postseismic response soon thereafter. 6. RSL change from corals during interseismic periods Our preferred interpretation of the coral records from Nias, Bangkaru, and southern Simeulue is that most sites experienced no In addition to recording details of sudden land-level changes change in 1843 (Fig. 2). Northwestern Nias, however, experienced associated with past earthquakes, the coral microatolls reveal that 266 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 7. Relative sea-level history (coral growth history) for site LAG, derived from slabs LAG-1, LAG-2, LAG-3B, and LAG-4. Elevations are plotted relative to the HLG just before the 16 February 1861 earthquake. Different colors represent data from different corals. As at site PBK, these fossil corals provide information on both the timing of the predecessor to the 2005 earthquake and the amount of uplift resulting therefrom. We estimate the 1861 uplift at LAG from the difference between the pre-diedown HLG on LAG-1 and the post- diedown HLS on LAG-3B, which is 28 cm, and then add 6 cm to account for the estimated erosion of the outer rim of LAG-1; this yields an estimate of ~34 cm. For the slab LAG-1 cross section, see Figure S75. interseismic deformation patterns can change suddenly in ways attempts to fit rates at certain sites, we find that only minor ad- that have not been appreciated generally. Here we document dra- justments are needed to yield rates at two sites for the later pre- matic, abrupt changes in subsidence rates in the century preceding 1861 period that satisfy the criteria for significance proposed by the 1861 earthquake that are broadly coherent across southern Meltzner et al. (2012). Specifically, the “average” rate at SMB over Simeulue and may also appear on northern Nias (Figs. 11 and 12). the period 1814e1861 (instead of 1819e1861) is 8.0 ± 1.7 mm/yr, Separately, we also document abrupt changes in the 20th-century fits the data well, and satisfies the Meltzner et al. (2012) criteria for rates of vertical deformation on Bangkaru Island that involve significance. Similarly, the “average” rate at LBJ over the period changes from interseismic subsidence to interseismic uplift and 1829e1861 (instead of 1838e1861) is 5.5 ± 2.5 mm/yr, fits the then back to subsidence (Fig. 13). data well, and satisfies the Meltzner et al. (2012) criteria for sig- nificance. Furthermore, as conceded by Meltzner et al. (2012), even fi 6.1. Observations on southern Simeulue and northern Nias in some cases where we cannot show a rate change to be signi - cant, observations might still be best explained by sudden rate Southern Simeulue experienced uniformly low rates of RSL changes that are real. change in the 18th century, but that changed considerably in the Although there is no consensus on global sea levels prior to the early 19th century. At all four southern Simeulue sites where we mid-19th century, and extrapolations from individual local sea- slabbed microatolls that died in 1861, the central upper surfaces of level histories to global eustatic sea level (even after correction those microatolls show little elevation gain as they grew, but all for glacial isostatic adjustment) are subject to biases resulting from began rising rapidly toward their outer perimeters in their final regional sea-level variability, a number of authors have concluded decades of growth (Figs. 4 and 11). Microatolls at PWG and AFL (on based on studies of proxy sea-level data that eustatic sea level fell northwestern Nias) hint at a similar pattern, though their records gradually from AD 1400 until 1850, at a rate of about 0.1 mm/yr, are less compelling. Fig. 11 shows time series of upward coral and that modern sea-level rise did not begin until after 1850 (Kemp fi growth, which serve as a proxy for RSL, for most of the sites. The et al., 2011, 2013; Cahill et al., 2015). These ndings suggest that most dramatic changes in the pre-1861 submergence rates (rates of sea-level change over the duration of our pre-1861 coral time series RSL rise) occurred in southern Simeulue, at SMB (0.8 ± 1.4 mm/yr, was negligible, and all documented changes in RSL from that period 1760 to 1819, increasing to 8.7 ± 1.9 mm/yr, 1819 to 1861); at UTG resulted from changes of the opposite sign in land level. Even if the (1.1 ± 1.0 mm/yr, 1757 to 1839, increasing to 7.0 ± 3.6 mm/yr, last few years of the coral time series coincided with a post-1850 1839 to 1861); and at LBJ (1.6 ± 0.8 mm/yr, 1738 to 1838, initiation of eustatic sea-level rise, the rate changes (except at our increasing to 6.1 ± 3.5 mm/yr, 1838 to 1861). northernmost site, SLR) all occurred at least a decade prior to 1850. Meltzner et al. (2012) proposed that only submergence rates At SMB the later, faster rate was established by HLG data points in based upon four or more HLG points, spanning three or more die- 1824 and 1840 (Figure S11), and at LBJ, the faster rate was deter- downs, can be considered significant; other rates, they noted, mined by HLG data points in 1841 and 1849 (Figure S26). should not be considered significant because of the brevity of those Even if the inferred eustatic sea-level history is wrong or not fl intervals and the potential for bias that could result from preser- re ective of regional sea-level history in the eastern Indian Ocean, vation peculiarities of a particular microatoll. Given these criteria, we have reason to preclude even regional changes in sea level as fi only the early slower rates at UTG and LBJ are significant signi cant contributors to the changes we observe in the corals. (Figures S21, S26); other pre-1861 rates are not statistically signif- Variations from global average eustatic sea-level trends can arise at ~ icant, strictly speaking. That said, if we reconsider our initial various timescales from climate anomalies such as the El Nino/ A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 267

Fig. 8. Maps showing uplift derived from coral microatolls, for earthquakes in AD 1422, 1843, 1861, and 2005. The 2005, 1861, and possibly 1422 earthquakes were similar to one another with regard to both extent and amount of uplift. In 1843, the uplift distribution is clearly different, and site PSN likely subsided. Sites where microatolls preclude statistically significant land-level change are indicated. For 2005, contours show uplift and subsidence in cm, updated from Briggs et al. (2006) and Meltzner et al. (2012). The circle color for each site is as in Fig. 2: yellow denotes sites with evidence for 1861 uplift, whereas red denotes sites with evidence for 1843 deformation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Southern Oscillation (ENSO) or Indian Ocean Dipole (IOD), and from supplementary text, but the method of analysis developed by persistent changes in currents or wind or wave direction. Because Meltzner et al. (2010) effectively mitigates any bias that IOD vari- Simeulue is only ~100 km long, any wind, wave, or circulation- ability may introduce into our estimates of subsidence rates. In driven changes in regional sea level (or any changes in eustatic contrast, the subsidence rates appear to increase around 1819 at sea level) should be seen synchronously at all sites on Simeulue. We SMB, around 1838 or 1839 merely 12 km to the southeast at UTG do see diedowns due to positive IOD events in the coral records, and and farther southeast at LBJ, and around 1849 only 8 km to the they tend to be synchronous at all sites, as discussed in the northwest at SLR. Although we cannot resolve the timing of rate 268 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Table 1 changes recorded by our southern Simeulue corals to better than a Coseismic deformation recorded by coral microatolls above southern Simeulue, decade, the 20e30 year difference in the timing of the change at Bangkaru, and Nias. various sites is unlikely to be an artifact of the analyses. In sum- Site ~1422 Uplift 1843 Uplift 1861 Uplift 2005 Uplift Notese mary, these changes are far more consistent with the spatiotem- (cm) (cm) (cm) (cm) poral scales over which tectonic deformation can vary, and decadal- BUN 66e77 ± 12 65e80 ± 4 a scale changes in regional sea level do not appear to significantly SLR No change 45 52 ± 8 affect our interpretations of subsidence rates. ± SMB No change 40 78 10 In Fig. 12, we invert the time series from Fig. 11 (i.e., we take the UTG No change 27 100 ± 6 LBJ No change 110 88 ± 6 negative of each time series), and we shift it vertically by 16 cm to PBK No change 30e35 ± 634± 7 b account for eustatic sea-level rise since the 20th century, in order to PWG No change 180 180 ± 19 determine land-level changes in the 18the19th centuries; hence, ± AFL 17 230 16 Fig. 12 is a vertical geodetic time series. We estimate rates of RSL PSN Subsidence? 150 ± 16 MZL No change 35 80 ± 23 change and corresponding rates of land-level change for various BWL Uplift? Uplift? 130 ± 16 c time periods in Table 2. LAG No change 34 ± 6 Betw. 30 and 60 d For 20th-century rates, we assume a sea-level rise of 2 mm/yr a From Meltzner et al. (2012). since 1925, for consistency with other results on Simeulue b The 1861 uplift at PBK was followed by 25 cm postseismic subsidence within (Meltzner et al., 2010, 2012), but we bear in mind two groups of <14 years. The 2005 uplift at PBK was followed by ~17 cm postseismic uplift, then studies that suggest sea-level rise may have been more compli- ~16 cm coseismic subsidence in 2010. cated. First, in modeling SSH trends in individual ocean basins, c A poorly dated uplift at BWL could have occurred in either 1843 or 1861. d The 2005 uplift estimated at LAG is based, in part, on interpolation between Jevrejeva et al. (2006) estimate nonlinear SSH trends over time uplift at nearby sites. within the Indian Ocean basin: they find that Indian Ocean sea e All errors should be considered as 2s. level, on average, rose by ~4 mm/yr from ~1930 to ~1947, by ~3 mm/

Fig. 9. Cross sections of (a) slab AFL-3 and (b) chiseled hand sample AFL-4, both from site AFL. The UeTh dates suggest both died in the historical 1843 earthquake, and both appear to have less than a few millimeters (less than 0.5 band) of erosion. AFL-3 experienced diedowns in mid-1817 and early 1833, among others, but AFL-4 did not because it was lower in the water. Both cross sections are plotted at the same scale; note the finer banding in AFL-4, a Goniastrea sp. coral, than in AFL-3, a Porites sp. coral. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 269

Fig. 10. Cross section through part of slab PSN-2 (slice a), from site PSN. Unlike at nearby site AFL, no coral mortality occurred at PSN in 1843, and unlike all other sites on Nias, no diedown occurred here in 1846. The lack of a diedown in mid-1846 suggests that subsidence occurred at the PSN site, either during the 1843 earthquake or as a postseismic response soon thereafter, thereby lowering the coral relative to sea level and inhibiting subsequent diedowns. For the full slab, see Figure S62. yr until ~1958, by ~2 mm/yr until ~1965, by ~1 mm/yr until ~1980, extrapolate the 20th-century rates back to 1861 and to consider the and by < 0.5 mm/yr since 1980. These rates are modeled primarily resulting implications. Simply extrapolating the modern subsi- from tide gauges in the northern, western, and central Indian dence rates (from Fig. 15) for the southern Simeulue and northern Ocean, and it is unclear how closely those trends track sea level off Nias sites, we can estimate what the elevation of each site would Sumatra. More recently, the altimetry records mentioned previ- have been following the 1861 uplift, if the interseismic rates did not ously suggest sea-level rise since 1993 ranged from ~2 mm/yr just change between then and 2005, if postseismic deformation northwest of Simeulue to ~3 mm/yr near Nias and the Batu Islands following 1861 was negligible, and if deformation associated with (Beckley et al., 2007; Hamlington et al., 2011). It is unclear how far the 1907 earthquake was also negligible. We can then compare this back in time those altimetry-based records can be extrapolated, but hypothetical elevation with the site's elevation before the 1861 they differ markedly from the basin-wide estimates of Jevrejeva earthquake (from Fig. 12); if all our assumptions are correct, the et al. (2006). In the absence of more reliable estimates for difference at each site would be an independent estimate of the Simeulue and Nias prior to 1993, a 2 mm/yr rate of sea-level rise 1861 uplift at the site. At two sites, this hypothetical 1861 uplift since 1925 is a reasonable simplification. (UTG: 110 ± 80 cm; MZL: 50 ± 40 cm) is within error of the 2005 Unlike the pre-1861 corals, the 20th-century corals on southern uplift locally, although the uncertainty at each of those sites is large. Simeulue and northern Nias do not record dramatic increases in At two other sites, however, this hypothetical 1861 uplift (SLR: tectonic subsidence rates mid-way through the interseismic period, 90 ± 30 cm; LBJ: 180 ± 20 cm) is larger than the uplift in 2005 if we assume a uniform rate of sea-level rise since at least 1945 (or (Table 1). These discrepancies suggest either that the 1861 uplifts at even if we were to assume accelerating rates of sea-level rise). This those sites were indeed larger, or that the average subsidence rates is the case even at LBJ, whose modern record extends back to 1945 in the first half of the interseismic period (including postseismic (Figs. 14 and 15; Table 2). Although the uncertainties on some of the subsidence after 1861 and coseismic subsidence in 1907) were rates should caution us against comparing the pre-1861 and pre- substantially lower than the rates measured by corals in the latter 2005 rates in too much detail, most pre-2005 rates appear to fall half of the 20th century. between the early (slow) pre-1861 rates and the late (fast) pre-1861 rates. The exception is at site LBJ, where the rate since 1945 appears to have been consistently as fast as or faster than at any point be- 6.2. Observations on Bangkaru Island, Banyak Islands tween 1738 and 1861 (Figs. 12 and 15; Table 2). Although the sudden interseismic rate changes (before 1861) Coral data from the Banyak Islands suggest two 20th-century caution against extrapolating observed rates beyond the respective reversals in the rate of RSL change. Modern microatoll PBK-4 on fi periods of observation, it is nonetheless an informative exercise to Bangkaru Island started growing in the 1950s and rst recorded a diedown in 1956. As is the case with almost every microatoll in our 270 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 11. Relative sea-level histories for the 18the19th centuries for (a) southern Simeulue, (b) northern Nias, and (c) northern Bangkaru. Different colors represent data from different sites. On southern Simeulue and possibly northern Nias, rates of relative sea-level rise were slow prior to 1819e1839 but much faster from then until 1861. Rates on Bangkaru were fairly constant from 1812 or earlier until 1861; relative sea level dropped as land rose during the 1861 earthquake, but much of that change was recovered by 1875, by which time steady relative sea-level rise had resumed at the site. In (b), the elevation of all data from site AFL has been systematically shifted by 50 cm to account for an inferred surveying error, discussed in the supplement, Text S9.2. For the uncorrected elevations, see Figure S83. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 271 study, the first few diedowns were successively higher on the coral. 6.3.1. Southern Simeulue But there was then an abrupt change, and for ~15 years, successive In order to test for along-strike variations in the locking depth diedowns were lower and lower. The trend then, just as suddenly, under Simeulue, we divided the model fault into two (or for one reversed again, and successive diedowns were higher (Fig. 13). The time snapshot, three) sections along strike, each with a different rates of RSL change we estimate from PBK-4 are 4.7 mm/yr RSL rise locking depth. In all iterations of the model for Simeulue, we fixed from 1956 to 1966, 5.8 mm/yr RSL fall from 1966 to 1981, and the coupling (locking) ratio along the portion of the fault shallower 4.4 mm/yr RSL rise from 1981 until 2005. than 18 km at 0.4; deeper than that, various patches along the fault The PBK-4 record is unique among our corals not only for the were assigned either as fully locked (back-slipping at the subduc- change from submergence to emergence and back, but also for the tion rate) or as creeping. The region of partial coupling above 18 km unparalleled precision in both time and elevation with which we depth is consistent with the seismicity recorded by Tilmann et al. can resolve these rates. PBK-4 grew consistently at the rapid rate of (2011): a zone of intense seismicity occurs updip of southeastern 25 mm/yr and recorded a diedown once every two years on Simeulue, at depths of 15e18 km below sea level, and is inferred to average. This allows us to pinpoint the timing of rate changes to coincide with the upward transition from unstable sliding (seismic within ±2 years (which is much better than we can do using typical behavior) to stable sliding (aseismic behavior). One sample geom- corals), and it allows us to determine rates, with lower errors and etry is shown in Figure S88; this example illustrates a model with a with statistical significance, over periods of a decade or slightly less locking depth change from 50 km in the northwest to 25 km in the (see details in supplementary text). Hence, we have high confi- southeast. dence that these rates are robust, and that the changes in rates are For each configuration of along-strike locking depth change, we both rapid (over no more than 2e4 years but possibly over a period calculated an along-strike surface uplift rate profile for a hypo- of months or less) and real. Adjusting for 20th-century sea-level thetical row of surface points located 110 km from the model rise at a rate of 2 mm/yr, we determine that, from the beginning trench. This distance corresponds to the approximate distance from of the record in 1956 until 1966, the site subsided at 2.7 ± 3.4 mm/ the trench (as defined by Bird (2003)) of the four southern yr; it then suddenly switched to þ7.8 ± 2.1 mm/yr of gradual tec- Simeulue coral sites for which we have interseismic rates tonic uplift for 15 years, before abruptly reverting (Table S4). These surface uplift rate profiles were calculated as to 2.4 ± 0.8 mm/yr of subsidence from 1981 until coseismic uplift deformation due to a dislocation in an elastic half space (Okada, in 2005 (Fig. 15; Table 2). 1985, 1992). If, as for Simeulue and Nias, we extrapolate the 2.7 ± 3.4 mm/ From Table 2,wedefined four time periods on Simeulue, and we yr of subsidence at Bangkaru from 1956 (Fig. 15) to 1890 (Fig. 12), attempted to model the interseismic subsidence rates at three sites there would be a 40 ± 20 cm discrepancy, with the extrapolated during each time period. From the fossil microatoll records, we elevation being higher than the site's actual elevation as deter- modeled the rates at sites SMB, UTG, and LBJ (Fig. 16), which are the mined from corals. This discrepancy could be explained if there was best constrained. We ignored rates from other sites, which have uplift at the site in 1907, or if there were additional periods of large uncertainties or other ambiguities in interpretation. We gradual uplift between 1890 and 1956 that were similar to divided the fossil microatoll records into the periods pre-1819 1966e1981. (going back to the beginning of the microatoll records), post-1839 (through 1861), and a transition period between 1819 and 1839; during this transition, site SMB had already switched to a faster 6.3. Modeling of interseismic deformation rate, but the other sites had not. From the modern (pre-2005) microatoll records, we modeled the rates at sites SLR, UTG, and LBJ We employed elastic dislocation modeling in an effort to explain (Fig. 16), which are the best constrained for that period. Sites SLR the various observed rate changes on Simeulue and Bangkaru. We and SMB are only 8 km apart. developed back-slip models (Savage,1983), incorporating a variably At this stage, it is important to define our goals for these dipping fault geometry that approximates the Slab 1.0 model modeling efforts. In the present paper, we simply want to deter- (Hayes et al., 2012) for this section of the megathrust (Figure S87), mine whether models exist that can explain the data: we wish to and which fits the depth of aligned seismicity beneath southeastern test whether the rates and rate changes we infer from the corals are Simeulue recorded by an ocean bottom seismic (OBS) array physically plausible and can be explained by reasonable conditions (Tilmann et al., 2011). The model slab extends to a depth of 100 km; on the underlying megathrust. We are not exploring an exhaustive all model depths will be expressed relative to sea level. We set of forward models, and we are not attempting to find the best assumed a subduction (convergence) rate of 40 mm/yr, based on possible model. In particular, we are exploring how along-strike the rate reported by McNeill and Henstock (2014) for this section of variations in the downdip limit of locking might explain the the subduction zone. along-strike variations in subsidence rates on southern Simeulue, We attempted to model the observed variations in interseismic but we are not exploring how those rate variations might be rates as consequences of spatiotemporal changes in the locking explained, instead, by along-strike variations in the updip limit of depth along the megathrust. Separately for Simeulue and for full locking or by along-strike variations in plate coupling. Such Bangkaru, we considered a range of plausible forward models. models could and should be tested along with a more rigorous Because the southern Simeulue sites are roughly equidistant from exploration of the model space. the trench but experienced different subsidence rates at different For each of the four periods, we determined the model (among times, we tried to explain the observations by modeling different those considered) that yields the best visual fit to coral observa- along-strike variations in the locking depth at different times. In tions during that period (Fig. 17). From these simple forward this manner, for any snapshot in time, we would be required to models, we conclude that one way (but probably not the only way) explain all the southern Simeulue rates with a single 3-dimensional to explain the observations is with: (a) locking down to 26 km locking pattern, but for a different decade or century, a different depth under Simeulue and slightly deeper locking immediately to locking pattern might be required to explain the data. The Bangkaru the southeast prior to 1819; (b) locking down to 50 km under site is sufficiently far from other sites that we modeled the rates Simeulue but down to only 25 km immediately to the southeast there independently, considering changes only over a 2- from 1839 to 1861; and (c) locking down to 28 km under northern dimensional profile perpendicular to the trench. Simeulue but down to 45 km under southern Simeulue in the 20th 272 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 12. Histories of interseismic subsidence (and where constrained, coseismic uplift) through the 18the19th centuries at sites on southern Simeulue, northern Nias, and Bangkaru Islands. Subsidence rates abruptly increased on southern Simeulue and possibly northern Nias in the decades prior to the 1861 earthquake, whereas rates were fairly uniform on Bangkaru before 1861 and after 1875. Shown in a geodetic reference frame, these rates and elevations have been inverted from the corresponding relative sea-level histories (Fig. 11), and the time series have been shifted vertically by 16 cm to account for eustatic sea-level rise since the 20th century, following Meltzner et al. (2010). Data constrain solid parts of the curves well; dashed portions are inferred or less reliable; queried portions are questionable. Interseismic subsidence rates (all negative, in mm/yr) are shown. Vertical dotted white lines mark the 1843 and 1861 earthquakes. The zero elevation at each site is defined as the site's elevation immediately prior to the 2004e2005 uplift. century. Viewed another way, the locking depth southeast of earlier (Fig. 1). Hence, these models do not inform us about simi- Simeulue remained at 25e30 km for the entire pre-1861 period, but larities or differences in locking depth across the barrier. We locking deepened substantially under Simeulue after 1819. Prior to consider this an effort worthy of future investigation. 2005, the pattern was different, with deeper locking under south- ern Simeulue and farther southeast. 6.3.2. Bangkaru For the 1819e1839 transition period, even with a more In an attempt to model the rate variations at Bangkaru (Fig. 18), complicated three-section model, we cannot do a good job of we used similar back-slip and elastic dislocation models as simultaneously fitting all three rates (Fig. 17). This calls into ques- described for Simeulue. Specifically, the same fault geometry and tion our interpretation of the time series. As discussed in Section subduction rate were incorporated into the model. However, 6.1, we cannot resolve the timing (or abruptness) of rate changes considering that the Bangkaru site is farther from the trench recorded by our southern Simeulue corals to better than about a (Table S4) and is therefore less sensitive to locking patterns near the decade. Perhaps the change was more gradual than we inferred at trench, we simplified the model by eliminating the shallow region site SMB, and perhaps the subsidence rate at SMB between 1819 of partial coupling and instead extended the fully locked patch all and 1839 was not quite as fast as 8.7 ± 1.9 mm/yr. the way to the trench. Lastly, we caution that no rates have been modeled from To begin, we produced a set of interseismic surface uplift rate northwest of the persistent barrier under Simeulue, discussed profiles for various downdip limits of locking, and we compared A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 273

Fig. 13. The modern microatoll from the PBK site has an unusual growth history. Relative sea level at the site was rising gradually as the coral began to grow in the 1950s, but, after ~1966, successive diedowns were lower and lower, indicating relative sea level was gradually falling. Around 1981, the trend reversed again, and relative sea level gradually rose until the 2005 earthquake. We infer these changes to result from changes in rates of vertical tectonic deformation. (a) Cross section of slab PBK-4 (slice a); for cross sections of two parallel slices, see Figure S34. (b) Relative sea-level history (coral growth history) for site PBK, derived from and plotted at the same vertical scale as slab PBK-4. The 6-cm diedownin late 1971 is a little larger than most and may reflect a pulse of more rapid uplift within the 15-year period of gradual uplift; if so, it would be at the limit of our resolution.

those profiles to the fossil and modern rates estimated at the therefore attempted to model the effects at the Bangkaru site of a Bangkaru site (Fig. 18b). Although the 1981e2005, 1956e1966, and SSE on the megathrust in Sumatra (Fig. 19). various 1751e1894 subsidence rates are similar to one another and Continuing in our use of the theory of elastic dislocations, we can be explained by a range of plausible locking depths, the modeled the surface displacements from a SSE as the superposition 1966e1981 uplift rate is an exception and is difficult to model with of (a) deformation from steady creep at depth and (b) deformation simple back-slip models. If the fault is locked down to a depth of from thrust slip at greater than the plate convergence rate on a 30 km below sea level, which happens to produce an uplift peak patch within the otherwise locked zone. precisely at the Bangkaru site, then the modeled rate barely over- To determine the appropriate steady-state deformation to use in laps the low end of the 2s error bars of the observed rate (Fig. 18b). our model, we simply took the profile from Fig. 18b that best fits the Otherwise, no locking depth can fit the 1966e1981 uplift rate if a 1981e2005 vertical deformation rate, e2.4 ± 0.8 mm/yr. We chose realistic subduction (convergence) rate is used. the 1981e2005 rate because (a) no SSE is inferred during that Although we do not show models in which we attempt to period, (b) that rate is tightly constrained and reliable, and (c) it explain the 1966e1981 uplift rate with variations in the updip limit agrees with the rates at the site for all other times during which no of locking or with variations in coupling, no alternatives are likely SSE is inferred. The 1981e2005 rate is reasonably well fit by a 43- to work. The profiles shown in Fig. 18b already assume 100% km locking depth (Figs. 18b and 19a); below 43 km, the fault is coupling down to the stated locking depths; any uniform deviation effectively freely slipping at 40 mm/yr, the full subduction rate from 100% coupling over those depth ranges would simply lower (McNeill and Henstock, 2014)(Fig. 19a). the amplitude of the profiles and make the misfit greater. And We next modeled a family of SSEs, spanning depths from 20 to because the Bangkaru site is so far from the trench, any realistic 43 km and slip rates from 46 to 340 mm/yr. To illustrate the variations in the updip limit of full locking are unlikely to have a method, the surface deformation profile associated with a slow-slip significant effect at Bangkaru. (For instance, if we were to assume patch between 30 and 43 km depth and a slip rate of 49 mm/yr is the coupling ratio along the portion of the fault shallower than shown in Fig. 19b. This is essentially the profile that would be 18 km is 0.4, as we assumed for Simeulue, then, at 135 km from the associated with a coseismic rupture with uniform slip of 735 mm, trench, the subsidence rate would be merely 0.4 mm/yr faster than but extended over a 15-year period. Adding the surface deforma- when we modeled the fault as fully locked to the trench.) Hence, it tion profiles that result from the SSE and the steady slip at depth appears that the 1966e1981 uplift rate at Bangkaru is simply too yields the surface deformation profile that would be realized during fast to be explained by standard back-slip models. the proposed SSE. Our preferred model (a slow-slip patch at Inspired by long-duration SSEs in southern Alaska that cause 30e43 km depth, slipping at 49 mm/yr, with ongoing steady slip sites to abruptly start uplifting, undergo sustained gradual uplift for below that) is shown in Fig. 19c, in comparison with the Bangkaru several years, and then abruptly cease uplifting (Fu and site uplift rate from 1966 to 1981. Although we have not explored Freymueller, 2013), we wondered whether a similar phenomenon an exhaustive set of plausible SSEs, this SSE (Fig. 19c) does a far might be responsible for the gradual uplift at Bangkaru. We better job of fitting the observed 1966e1981 uplift rate than the 274 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Table 2 earthquake was an enigmatic event, producing strong shaking on Interseismic deformation rates recorded by coral microatolls above southern Nias and a tsunami that devastated the island's main town on the Simeulue, Bangkaru, and Nias. east coast. Our evidence suggests that land-level deformation Site RSL risea (mm/yr) Uplifta,b (mm/yr) Valid years (AD) Notes occurred primarily in northwestern Nias, at sites AFL and PSN Fossil rates (Fig. 8). It remains unclear whether this uplift resulted from rupture SLR 0.9 ± 1.6 0.9 ± 1.6 1800 to 1849 along the megathrust or from displacement along a nearby upper- 4.8 ± 6.7 4.8 ± 6.7 1849 to 1861 plate splay fault. ± ± SMB 0.8 1.4 0.8 1.4 1760 to 1819 The 1861 earthquake resembled that in 2005 in many ways. 8.7 ± 1.9 8.7 ± 1.9 1819 to 1861 UTG 1.1 ± 1.0 1.1 ± 1.0 1757 to 1839 Prior to the present study, Newcomb and McCann (1987) had 7.0 ± 3.6 7.0 ± 3.6 1839 to 1861 already suggested that the 1861 rupture involved the portion of LBJ 1.6 ± 0.8 1.6 ± 0.8 1738 to 1838 the megathrust between the Batu Islands and the Banyak Islands, ± ± 6.1 3.5 6.1 3.5 1838 to 1861 essentially the southern two-thirds of the 2005 rupture. The coral PBK 2.2 ± 0.7 2.2 ± 0.7 1751 to 1861 c 2.5 ± 1.6 2.5 ± 1.6 1812 to 1861 c records we present show that rupture in 1861 extended north- 4.6 ± 2.9 4.6 ± 2.9 1833 to 1861 c ward to the northern limit of rupture in 2005 (Fig. 8). At sites PBK 3.0 ± 4.2 3.0 ± 4.2 1875 to 1894 and LAG, where 1861 uplift is well determined, it was indistin- PWG 3.5 ± 2.1 3.5 ± 2.1 1791 to 1829 guishable from estimates of uplift at those locations in 2005 ± ± 8.5 2.5 8.5 2.5 1829 to 1861 (Fig. 8; Table 1): at PBK, the 30e35 cm of uplift in 1861 is similar to AFL 7.0 ± 5.0 7.0 ± 5.0 1817 to 1833 10.7 ± 8.0 10.7 ± 8.0 1833 to 1843 d an estimated ~34 cm of uplift at the site in 2005, while at LAG, the PSN 7.2 ± 5.0 7.2 ± 5.0 1814 to 1830 1861 uplift of ~34 cm compares to an estimated 30e60 cm of MZL 4.4 ± 2.1 4.4 ± 2.1 1823 to 1861 uplift at the site in 2005 (see details in supplementary text). At ± ± LAG 2.2 1.8 2.2 1.8 1817 to 1861 sites SLR and PWG, if 1861 uplift did not substantially exceed our 4.3 ± 2.4 4.3 ± 2.4 1866 to 1899 Modern rates minimum estimates, then the 2005 uplifts mimicked those in SLR 6.2 ± 2.2 4.2 ± 2.2 1961 to 1997 1861. At site LBJ, if 1861 uplift did not substantially exceed our UTG 6.2 ± 5.3 4.2 ± 5.3 1982 to 1997 minimum estimate, then the two uplifts there did not differ by LBJ 10.2 ± 1.5 8.2 ± 1.5 1945 to 1997 more than ~25% (Fig. 8; Table 1). ± ± PBK 4.7 3.4 2.7 3.4 1956 to 1966 a The similar estimates for coseismic uplift in 1861 and 2005 at 5.8 ± 2.1 7.8 ± 2.1 1966 to 1981 a 4.4 ± 0.8 2.4 ± 0.8 1981 to 2005 a PBK notwithstanding, the deformation that followed the two AFL 9.9 ± 2.0 7.9 ± 2.0 1956 to 1997 earthquakes is dramatically different. The 1861 uplift at PBK was MZL 4.5 ± 3.0 2.5 ± 3.0 1970 to 1997 followed by 25 cm of postseismic subsidence within 14 years or a Rate errors were determined according to Meltzner et al. (2012), except for PBK less, with no diedowns recorded at the site between 1861 and 1877 modern rates. For all rates at PBK between 1956 and 2005, errors were determined (Fig. 6). (The corals do not tightly constrain how quickly or linearly by linear regression. All rate errors in this table should be considered as 2s. this postseismic subsidence occurred; all we know is that it b For fossil rates, sea level is assumed to be stable, hence the uplift is simply the occurred quickly enough that the coral experienced unconstrained negative of RSL rise. For modern rates, we assume 2 mm/yr sea-level rise; this must be subtracted out to calculate land-level change. upward growth during that period.) In contrast, our ongoing c The 1751e1861 rate at PBK is an average over that period, but we cannot measurements of RSL change at site PBK, as well as a gradual die- preclude rate changes prior to 1812. The 1751e1861 and 1812e1861 rates at PBK down recorded by PBK-1 (Figure S29) suggest ~17 cm of post- e were determined from microatolls at subsite PBK-B. The 1833 1861 rate at PBK was seismic uplift occurred between May 2005 and January 2009, determined independently from a microatoll at subsite PBK-C, 1.1 km away. d followed by ~16 cm of coseismic subsidence in a M The 1833e1843 rate at AFL does not include coseismic uplift in 1843. W 7.8 earthquake in April 2010. Notably, just after the April 2010 earthquake, the site was back to nearly the exact elevation where it had been just after the March 2005 earthquake, a curious feature observed at other standard back-slip model could alone (Fig. 18b). There clearly will sites in the Banyak Islands. Preliminary GPS data from the Sumatran be a tradeoff between slip rate and slip area, so more data would be GPS Array (SuGAr) suggest Bangkaru subsided ~2 cm between needed to constrain a truly “best” SSE model; nevertheless, our January 2012 and January 2013, although the available time series is model indicates that a SSE is a plausible explanation for the ob- too short and has too many gaps to allow for a robust analysis. servations, while variations in the extent of the locked region are Nonetheless, the behavior of Bangkaru following 1861 appears not. more like its behavior following the 2010 earthquake than Lastly, we recall that rates of sea-level change prior to 1992 are following the 2005 earthquake. We propose that the patches of the not well determined for the eastern Indian Ocean. If, contrary to megathrust that slipped discretely during the 2005 earthquake, the our assumption of (and correction for) 2 mm/yr sea-level rise 2005 postseismic period, and the April 2010 earthquake ruptured between 1966 and 1981, there was instead zero sea-level change in tandem during the 1861 earthquake, and we hypothesize that over that period, the uncorrected 1966e1981 uplift rate (5.8 mm/ the site will continue to subside rapidly following the April 2010 yr) would be easier to model with simple back-slip, but that uplift earthquake, as it did after 1861, recovering much of the 2005e2010 would still represent a marked and abrupt deviation from the uplift in the decades to come. subsidence that preceded and followed. In that sense, the rate Our data do not provide additional information on the 1907 changes on Bangkaru would still resemble those during the SSE in rupture, because no corals were found on the reef flats of southern southern Alaska. Simeulue, Bangkaru, or Nias that were living at the time. At most Simeulue and Nias sites just prior to 1907, the reef flats were likely 7. Discussion still elevated above the base of the intertidal zone following the 1861 uplift, preventing coral colonies from establishing. Coseismic 7.1. Earthquake history subsidence in 1907, as reported in nautical guidebooks (Great Britain Hydrographic Department, 1934) (see details in the Ap- Our analyses of coral microatolls from Nias, Bangkaru, and pendix), combined with ongoing interseismic subsidence and sea- southern Simeulue reveal details of tectonic deformation during level rise, may have allowed corals to finally recolonize the reef flats two earthquakes in the region in the 19th century. The 1843 around 1940 at LBJ and slightly later at other sites. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 275

Fig. 14. Relative sea-level history for the 20th century for (a) southern Simeulue, (b) northern Nias, and (c) northern Bangkaru. Different colors represent data from different sites. Abrupt changes in the rates of relative sea-level change were seen on Bangkaru, but not elsewhere.

Corals provide limited information about an earlier rupture of of the ~1422 uplift to the south or east are known, however, pre- the NiaseSouthern Simeulue patch, in the 15th century. At the BUN cluding further assessment of the ~1422 rupture. Additional uplifts site, along the southwestern coast of Simeulue, a 66e77 cm pre- along the NiaseSouthern Simeulue patch may have occurred in the historic uplift around AD 1422 was similar to the 65e80 cm uplift at 16th, 17th, and/or 18th centuries, but details of any such events the site in 2005 (Meltzner et al., 2012)(Fig. 8). In further likeness to have yet to be resolved. one another, both the ~1422 and 2005 ruptures terminated under To the north of the 2005 rupture, on northern Simeulue, two the northwestern half of Simeulue, with little to no land-level paleoseismic events (~1394 and ~1450) are reasonably well con- change at sites on northern Simeulue in either event. No details strained, but the uplift distributions in each event and in 2004 are 276 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 16. Rates of interseismic land-level change during various time periods on southern Simeulue. The error ellipses have a fixed width for visualization purposes only.

rupture and the amount of uplift at some, if not many, sites. The Fig. 15. Histories of interseismic vertical deformation preceding the 2005 earthquake ~1422 rupture may have also been similar to 2005. Although other at sites on southern Simeulue, northern Nias, and Bangkaru Islands. Although un- earthquakes, not similar to 2005, have also affected this region certainties on some rates are large, the pre-2005 rates on Simeulue and Nias generally (including in 1843, either on a portion of the megathrust or on a differ from the pre-1861 rates shown in Fig. 12. On Bangkaru, interseismic subsidence before 1966 switched to gradual uplift until 1981, and then reverted to subsidence splay fault; in 1907, likely on the shallow part of the megathrust, through 2005. Shown in a geodetic reference frame, these rates and elevations have updip of the 2005 patch (Kanamori et al., 2010); and in April 2010), been inverted from the corresponding relative sea-level histories (Fig. 14), and the time any similarities among the largest ruptures of the NiaseSouthern series have been adjusted by 2 mm/yr to account for 20th-century eustatic sea-level Simeulue portion of the megathrust are noteworthy. rise. Interseismic uplift (positive) or subsidence (negative) rates are given in mm/yr. The NiaseSouthern Simeulue section of the megathrust off Vertical dotted white line marks the 2005 earthquake. The zero elevation at each site is taken as the site's elevation immediately prior to the 2004e2005 uplift, as defined by Sumatra is also unique in that it appears to be the only patch that is the HLG in late 2004. Some pre-1997 rates project to 2004 elevations that are lower tightly bounded by two persistent barriers to rupture (Natawidjaja than zero either because the site rose in an earthquake in 2002, or because the coral et al., 2006; Meltzner et al., 2012; Philibosian et al., 2014)(Fig. 1). had not yet grown back up to its HLS following the late 1997 IOD diedown. We propose that these two characteristics of this portion of the megathrust are not unrelated: by confining ruptures to a comparatively short section of the megathrust, the barriers may dramatically different from one another (Meltzner et al., 2010). lead to a higher degree of similarity between successive ruptures South of the 2005 rupture on the Mentawai patch, each of the here than elsewhere. In this sense, segmentation and rupture past four supercycles involved a unique rupture sequence in repeatability would be related, as has been proposed along certain terms of the groupings of asperities that ruptured in individual strike-slip faults (e.g., Rockwell et al., 2001). Perhaps because the earthquakes (Sieh et al., 2008, 2014). Hence, north of the north- segment boundaries are such persistent rupture barriers and the ern barrier and south of the southern barrier of Meltzner et al. intervening stretch of the megathrust has comparatively little (2012), there is a high degree of variability among ruptures. The heterogeneity in frictional properties, many ruptures tend to extent to which similarities might exist among ruptures of the extend to the full distance between the barriers. A NiaseSouthern Simeulue patch is unique along the Sunda proposed by Jacob et al. (2014) in the downgoing slab between megathrust. Simeulue and Nias may divide the deep portion of the Although the ruptures in 1861 and 2005 were not identical to NiaseSouthern Simeulue patch into two primary asperities; one another or perfectly “characteristic,” they bore remarkable nonetheless, both asperities ruptured in 2005 (Briggs et al., 2006) similarities to one another, in terms of the along-strike extent of and apparently in 1861. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 277

Fig. 17. Back-slip (elastic dislocation) models suggest that most of the spatiotemporal variations in interseismic subsidence rates on southern Simeulue can be explained by along- strike variations in the locking depth along the megathrust. The upper panels show profiles of predicted uplift (or subsidence) rates along a hypothetical row of surface points located 110 km from the trench (the approximate distance from the trench of our sites), as a consequence of along-strike variations in the downdip limit of locking. In each of the “Pre-1819,”“1839e1861,” and “Pre-2005” upper panels, several models are shown: on one side of the boundary (the side that happens to have deeper locking) the locking depth is fixed, but on the other side, along-strike uplift-rate profiles are shown for a selection of locking depths to illustrate the effect of varying the locking depth. Subsidence rates estimated from corals are overlain on these model-predicted profiles. For each time period, the preferred model profile is indicated with a thicker line. Back-slip models have more difficulty explaining the spatial distribution of subsidence rates between 1819 and 1839, a transition period during which SMB had already experienced an increase in the sub- sidence rate, but the other sites had not. Even with a three-patch model, it is difficult to simultaneously fit all three rates during the 1819e1839 transition period. (See text for further discussion.) For all time periods, the assumed subduction (convergence) rate is 40 mm/yr. The lower panels show the surface projections of the partially (light blue) and fully (yellow) locked patches in the preferred model for each time period; downdip limits of full locking (depths in km) are labeled. Along-strike boundaries between patches with different locking depths are arbitrarily located to maximize the fit to the data. The coupling ratio along the portion of the fault shallower than 18 km (light blue patches) is fixed at 0.4 in all models. Arrows indicate observed subsidence rates; the actual years over which each rate is determined is indicated in Fig. 16. Isobaths along the megathrust from Hayes et al. (2012) are shown at 20-km intervals. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

7.2. Temporal variations in interseismic rates indistinguishable from zero change) between 1980 and presum- ably 2002 (Meltzner et al., 2012) (uncertainties are 2s). The coral microatolls provide unprecedented records of abrupt Philibosian et al. (2014) documented changes in subsidence changes in rates of interseismic vertical deformation. Two separate rates in the Mentawai Islands that coincided with the 1797 earth- data sets within our study are particularly interesting. One data quake there, but our rate changes do not appear to relate to nearby set allows us to resolve an “event” that occurs over the timescale of earthquakes. Even regional earthquakes such as the 1833 Mentawai just more than a decade. The other offers insight into phenomena earthquake cannot explain rate changes observed at the SMB site, that take place over decades to centuries. which occurred more than a decade earlier. Philibosian et al. (2014) On southern Simeulue and perhaps on Nias, the interseismic also observed a discrepancy between the average coupling in the subsidence rates increased suddenly, decades before the 1861 Mentawai Islands from 1950 to 2000, as suggested by corals, and earthquake. The post-increase rate can be a factor of 4e10 greater the coupling suggested by horizontal campaign GPS data collected than the pre-increase rate at the site (Fig. 12; Table 2), and the lack over 1991e2001; one way to explain this discrepancy would be of synchrony at sites merely ~8 km apart precludes the possibility with an increase in coupling around 1990, at a time when no large that these changes resulted from regional or global sea-level fluc- earthquakes occurred. If this proposed change in the Mentawai tuations. Nevertheless, as pronounced as the pre-1861 changes are, Islands is real, perhaps it is analogous to the changes we observe on no rate changes are seen at those sites in the 40e60 years prior to southern Simeulue. the 2005 earthquake. One way to explain the rate changes on Simeulue is with In addition to these examples where subsidence rates changes in the locking depth in both time and space along the remained steady or increased in the decades before an earth- megathrust. Our modeling results suggest that some parts of the quake, a modern microatoll at the Pulau Penyu (PPY) site in fault that were locked only down to ~26 km depth for decades central Simeulue (Fig. 2) highlighted at least one example in during the 18th and early 19th centuries may have become locked which subsidence may have suddenly slowed or stopped in the down to ~50 km in the decades before the 1861 earthquake (Fig.17). decades before an earthquake. Around 1980, the rate there Neighboring parts of the fault that were also locked only to changed abruptly, from 3.9 ± 1.7 mm/yr (subsidence) between 25e30 km depth during much of the 18th and 19th centuries were 1932 and 1980, to þ0.3 ± 3.6 mm/yr (very slow uplift, but locked down to ~45 km for the second half of the 20th century 278 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Fig. 18. Back-slip models, incorporating a range of plausible locking depths, can readily explain the subsidence rates at Bangkaru during the 18the19th centuries, from 1956 to 1966, and from 1981 to 2005, but no standard back-slip model does a good job of fitting the 1966e1981 uplift rate; the observed rate is simply too fast. (a) Rates of interseismic land-level change during the 20th century on Bangkaru. The error ellipses have a fixed width for visualization purposes only. (b) Model predictions of surface uplift and subsidence are plotted as a function of distance from the trench, for various choices of the downdip limit of locking (LD, locking depth). Superimposed on these model predictions are observed vertical deformation rates at the PBK site on Bangkaru, for the 18the19th centuries (left) and for the 20the21st centuries (right). The assumed subduction (convergence) rate is 40 mm/yr, and for simplicity full coupling is assumed all the way to the trench; nonetheless, the state of coupling near the trench has little effect at the PBK site, 135 km from the trench.

Fig. 19. Back-slip models, modified to incorporate a slow slip event (SSE; slip at rates exceeding the plate convergence rate) on a portion of the otherwise locked fault zone, can explain the steady but rapid uplift at PBK between 1966 and 1981. Trench-normal model profiles of uplift and subsidence rates, under the specified conditions, are shown in the upper panels. (a) Steady deformation (no SSE): the fault is fully locked above 43 km and freely slipping at the full subduction rate below 43 km. Superimposed on the model profile is the subsidence rate at the PBK site between 1981 and 2005, a period during which no SSE is inferred. (b) Deformation from slow slip at 49 mm/yr between 30 and 43 km depth. Ongoing steady deformation below 43 km is not considered in (b), so this profile is hypothetical only and should not match real-world observations (no data are overlain). (c) SSE between 30 and 43 km depth, with ongoing steady slip below that; this is the sum at each point along the profile of curves (a) and (b). This profile is intended to match real-world observations during the SSE; hence, we superimpose the uplift rate at the PBK site between 1966 and 1981 for comparison. Lower panels depict a simplified geometry of the model fault: the fault is normally locked above 43 km depth, but between 1966 and 1981 a SSE occurs between 30 and 43 km, with slip at ~49 mm/yr. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 279

(Fig. 17). If the locking depth variations we have modeled explain 8. Conclusions the rate changes on Simeulue, then they may be similar to abrupt changes in the width of the locked region that have been observed Coral microatolls reveal histories of interseismic strain accu- along the Alaska subduction zone near Lower Cook Inlet in south- mulation and coseismic strain release along the patch of the Sunda ern Alaska (Freymueller et al., 2014; Li et al., 2014). megathrust that ruptured in 2005. Combined with historical in- Farther from the trench, on Bangkaru, dramatic and sudden formation, these records provide abundant evidence for two changes occurred in the half-century preceding the 2005 earth- earthquakes, and suggestions of a third, that are similar to one quake, with linear interseismic subsidence switching to linear another. The extent of the 1861 rupture mimics that in 2005, and interseismic uplift, and then back to linear interseismic subsidence at the sites where the 1861 uplift is best resolved, it resembles that 15 years later (Fig. 15; Table 2). The rate changes on Bangkaru are of 2005. At one site near the northwestern end of the 2005 abrupt, occurring over a span of no more than 2e4 years (our limit rupture, uplift in a prehistoric event around AD 1422 was similar of resolution) but possibly over a period of months or less. to that in 2005, and the northwestern terminus of the 1422 The rate changes observed on Bangkaru also bear remarkable earthquake matched that in 2005. The higher degree of similarity resemblance to rate changes observed over the Alaska subduction between these ruptures, in contrast to ruptures on adjacent sec- zone. Near Upper Cook Inlet, 300 km north-northeast of Lower tions of the fault, may be related to the position of the patch be- Cook Inlet and the region where Freymueller et al. (2014) docu- tween two persistent barriers. Nonetheless, earthquakes that are mented changes in the width of the locked region, Fu and not near-repeats of 1861 or 2005 also occur between the central Freymueller (2013) analyzed and presented GPS time series that Simeulue and Batu Islands barriers: examples include the 1907 are consistent with long-duration SSEs downdip of the main and April 2010 ruptures of comparatively small portions of the asperity that ruptured in the 1964 Alaska earthquake. Specifically, megathrust, and the enigmatic 1843 earthquake that may have they suggest that the data capture the last part of a SSE from 1998 occurred either on the megathrust or on an upper-plate splay until 2001, and the first parts of another SSE from 2009 through the fault. end of the time series in late 2012. The vertical deformation rates in The coral microatolls also reveal changes that occur along the the Alaskan time series published by Fu and Freymueller (2013) are megathrust during what is commonly considered to be the roughly 10e15 mm/yr uplift during SSEs and roughly 1e2 mm/yr “interseismic” period, and they suggest that the range of slow slip uplift for the period between SSEs. The difference between the behavior is broader than previously appreciated. Although inter- average rate during the SSEs and the rate otherwise, ~11 mm/yr, is seismic vertical deformation rates (uplift or subsidence) may be nearly identical to the difference observed at Bangkaru. Uplift at linear for decades or even a century, the rate at any site may shift Bangkaru proceeded at 7.8 ± 2.1 mm/yr from 1966 to 1981, whereas abruptly to a new rate, and remain fixed at the new rate for decades subsidence rates estimated at the site over most other time periods more. Every southern Simeulue site experienced a sudden increase are indistinguishable from one another: among the more well in its subsidence rate in the decades before 1861, but the pre-2005 resolved rates at the site are 2.4 ± 0.8 mm/yr, 1981 to 2005; rates saw no increase and in general matched none of the rates e2.7 ± 3.4 mm/yr, 1956 to 1966; and 2.5 ± 1.6 mm/yr over 1812 to observed before 1861. In addition to these examples where subsi- 1861 or e2.2 ± 0.7 mm/yr over 1751 to 1861 (Table 2). dence rates remained steady or increased in the decades before an We propose that the Bangkaru and Upper Cook Inlet (Alaska) earthquake, earlier work in central Simeulue highlighted at least rate changes can be explained by the same phenomenon, likely a one example in which subsidence slowed or stopped suddenly in SSE. If so, the absolute difference in the rates (the generally higher the decades before an earthquake. The changes on Simeulue might rates of uplift in the Alaskan data set) can be explained by high be explained by abrupt shifts in the downdip limit of locking along rates of isostatic uplift due to deglaciation (following the Little Ice the megathrust; abrupt rate changes observed on Bangkaru can be Age) in southern Alaska (Larsen et al., 2005; J. T. Freymueller best explained by a 15-year-long slow slip event. In general, the (University of Alaska, Fairbanks), written communication, 2014), coral records suggest that our observations and understanding of and by the position of each observation site relative to the location fault behavior between earthquakes are still far from complete. of the SSE on the underlying interface. Our modeling results show They call into question the very meanings of an interseismic rate or that a SSE is a plausible explanation for the rate changes on an interseismic coupling pattern determined from a modern Bangkaru (Fig. 19). Unlike on Simeulue, however, changes in the geodetic network over a subduction zone, when that rate or width of the locked region alone cannot explain the rate changes coupling pattern may reflect motion and deformation during only on Bangkaru (Fig. 18). the most recent few decades. Although the modeling results presented here are rather crude (reflecting, in part, the limited resolution of the data), several Acknowledgments robust observations can be made. First, the locking depth, or the pattern of coupling along the megathrust under Simeulue, is not We thank D. Prayudi, I. Suprihanto, and J. Galetzka for field stationary in time, either over the course of one earthquake cycle, support; S. Martin and C. Vita-Finzi for help locating historical or from one earthquake cycle to the next. Indeed, the variations documents; and J. Freymueller, F. Taylor, R. Witter, and an anony- over decades or centuries can be dramatic. Second, standard back- mous reviewer for careful reviews and thoughtful suggestions that slip models cannot explain observed steady uplift rates at Bangkaru substantially improved this manuscript. This work has been sup- between 1966 and 1981. Third, the vertical deformation rates at ported by NSF grant EAR-0538333 (to K.S.); by MOST grants 102- Bangkaru are much better explained by a SSE under the island. 2116-M-002-016 and 103-2119-M-002-022 and NTU grant Although perhaps coincidental, one plausible hypothetical SSE 101R7625 (to C.C.S.); by NRF Fellowship NRF-NRFF2010-064 (to under Bangkaru (our preferred model among those tested) occurs E.M.H.); by LIPI (Indonesian Institute of Science) and RUTI (Inter- within the range of SSE depths reported along other subduction national Joint Research Program of the Indonesian Ministry of zones (Gomberg and The Cascadia 2007 and Beyond Working Research and Technology); by the Gordon and Betty Moore Foun- Group, 2010; Peng and Gomberg, 2010) and involves a slip rate dation; and by the Earth Observatory of Singapore and the National (~49 mm/yr) that is similar to that modeled during a long-duration Research Foundation (NRF) Singapore and the Singapore Ministry SSE under Upper Cook Inlet in southern Alaska between 2009 and of Education (MOE) under the Research Centres of Excellence at least 2012 (Fu and Freymueller, 2013). initiative. This is Earth Observatory of Singapore contribution 66. 280 A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281

Full-resolution coral slab x-ray mosaics for all of the slabs in this Hu, Y., Wang, K., He, J., Klotz, J., Khazaradze, G., 2004. Three-dimensional visco- fi paper are available from the corresponding author. We dedicate elastic nite element model for postseismic deformation of the great 1960 Chile earthquake. J. Geophys. Res. 109, B12403. http://dx.doi.org/10.1029/ this paper to the memory of Adi Rahman Putra, who gave his life in 2004JB003163. the pursuit of knowledge that might someday save the lives of Jacob, J., Dyment, J., Yatheesh, V., 2014. Revisiting the structure, age, and evolution others. of the Wharton Basin to better understand subduction under Indonesia. J. Geophys. Res. 119, 169e190. http://dx.doi.org/10.1002/2013JB010285. Jevrejeva, S., Grinsted, A., Moore, J.C., Holgate, S., 2006. Nonlinear trends and multiyear cycles in sea level records. J. Geophys. Res. 111, C09012. http://dx.doi. Appendix A. Supplementary data org/10.1029/2005JC003229. Junghuhn, F., 1845. Chronologisch overzigt der aardbevingen en uitbarstingen van Supplementary data related to this article can be found at http:// vulkanen in Neerland^ ’s lndie€ (in vergelijkende zamenstelling met elkander). Tijdschr. Neerl.^ Indie€ 7 (1), 30e68. Available at: http://books.google.com/books? dx.doi.org/10.1016/j.quascirev.2015.06.003. id¼FRMwAAAAYAAJ&pg¼PA30. Kanamori, H., Rivera, L., Lee, W.H.K., 2010. Historical seismograms for unravelling a mysterious earthquake: the 1907 Sumatra earthquake. Geophys. J. Int. 183, References 358e374. http://dx.doi.org/10.1111/j.1365-246X.2010.04731.x. Kemp, A.C., Horton, B.P., Donnelly, J.P., Mann, M.E., Vermeer, M., Rahmstorf, S., 2011. Abram, N.J., Gagan, M.K., Cole, J.E., Hantoro, W.S., Mudelsee, M., 2008. Recent Climate related sea-level variations over the past two millennia. Proc. Natl. intensification of tropical climate variability in the Indian Ocean. Nat. Geosci. 1, Acad. Sci. U. S. A. 108, 11017e11022. http://dx.doi.org/10.1073/pnas.1015619108. 849e853. http://dx.doi.org/10.1038/ngeo357. Kemp, A.C., Horton, B.P., Vane, C.H., Bernhardt, C.E., Corbett, D.R., Engelhart, S.E., Baird Smith, R., 1845. Register of Indian and Asiatic earthquakes for the year 1843. Anisfeld, S.C., Parnell, A.C., Cahill, N., 2013. Sea-level change during the last J. Asiat. Soc. Bengal 14 (2), 604e622. Available at: http://books.google.com/ 2500 years in New Jersey, USA. Quat. Sci. Rev. 81, 90e104. http://dx.doi.org/10. books?id¼L0wyAQAAMAAJ&pg¼PA604 https://archive.org/details/ 1016/j.quascirev.2013.09.024. journalofasiatic142asia. Konca, A.O., Avouac, J.-P., Sladen, A., Meltzner, A.J., Sieh, K., Fang, P., Li, Z., Galetzka, J., Beckley, B.D., Lemoine, F.G., Luthcke, S.B., Ray, R.D., Zelensky, N.P., 2007. Genrich, J., Chlieh, M., Natawidjaja, D.H., Bock, Y., Fielding, E.J., Ji, C., A reassessment of global and regional mean sea level trends from TOPEX and Helmberger, D.V., 2008. Partial rupture of a locked patch of the Sumatra Jason-1 altimetry based on revised reference frame and orbits. Geophys. Res. megathrust during the 2007 earthquake sequence. Nature 456, 631e635. Lett. 34, L14608. http://dx.doi.org/10.1029/2007GL030002. http://dx.doi.org/10.1038/nature07572. Beroza, G.C., Ide, S., 2009. Deep tremors and slow quakes. Science 324, 1025e1026. Larsen, C.F., Motyka, R.J., Freymueller, J.T., Echelmeyer, K.A., Ivins, E.R., 2005. Rapid http://dx.doi.org/10.1126/science.1171231. viscoelastic uplift in southeast Alaska caused by posteLittle Ice Age glacial Bird, P., 2003. An updated digital model of plate boundaries. Geochem. Geophys. retreat. Earth Planet. Sci. Lett. 237, 548e560. http://dx.doi.org/10.1016/j.epsl. Geosyst. 4, 1027. http://dx.doi.org/10.1029/2001GC000252. 2005.06.032. Briggs, R.W., Sieh, K., Meltzner, A.J., Natawidjaja, D., Galetzka, J., Suwargadi, B., Li, S., Freymueller, J.T., McCaffrey, R., 15e19 December 2014. Time-dependent var- Hsu, Y.-j, Simons, M., Hananto, N., Suprihanto, I., Prayudi, D., Avouac, J.-P., iations of slow slip events in Lower Cook Inlet of the Alaska-Aleutian subduc- Prawirodirdjo, L., Bock, Y., 2006. Deformation and slip along the Sunda mega- tion zone. Abstract S53C-4511 presented at 2014 Fall Meeting. American thrust in the great 2005 NiaseSimeulue earthquake. Science 311, 1897e1901. Geophysical Union, San Francisco, CA. http://dx.doi.org/10.1126/science.1122602. Mavrommatis, A.P., Segall, P., Johnson, K.M., 2014. A decadal-scale deformation Brown, B.E., Clarke, K.R., Warwick, R.M., 2002. Serial patterns of biodiversity change transient prior to the 2011 MW 9.0 Tohoku-oki earthquake. Geophys. Res. Lett. in corals across shallow reef flats in Ko Phuket, Thailand, due to the effects of 41, 4486e4494. http://dx.doi.org/10.1002/2014GL060139. local (sedimentation) and regional (climatic) perturbations. Mar. Biol. 141, McNeill, L.C., Henstock, T.J., 2014. Forearc structure and morphology along the 21e29. http://dx.doi.org/10.1007/s00227-002-0810-0. SumatraeAndaman subduction zone. Tectonics 33, 112e134. http://dx.doi.org/ Bürgmann, R., Thatcher, W., 2013. Space geodesy: a revolution in crustal defor- 10.1002/2012TC003264. mation measurements of tectonic processes. In: Bickford, M.E. (Ed.), The Web of Melbourne, T.I., Webb, F.H., Stock, J.M., Reigber, C., 2002. Rapid postseismic tran- Geological Sciences: Advances, Impacts, and Interactions, Geol. Soc. Am. Spec. sients in subduction zones from continuous GPS. J. Geophys. Res. 107, 2241. Pap. 500, pp. 397e430. http://dx.doi.org/10.1130/2013.2500(12). http://dx.doi.org/10.1029/2001JB000555. Cahill, N., Kemp, A.C., Horton, B.P., Parnell, A.C., 2015. Modeling sea-level change Meltzner, A.J., Woodroffe, C.D., 2015. Coral microatolls. In: Shennan, I., Long, A.J., using errors-in-variables integrated Gaussian processes. Ann. Appl. Statist. 9 (2). Horton, B.P. (Eds.), Handbook of Sea-Level Research. John Wiley & Sons, Ltd., Chambers, D.P., Tapley, B.D., Stewart, R.H., 1999. Anomalous warming in the Indian Chichester, UK, pp. 125e145. http://dx.doi.org/10.1002/9781118452547.ch8. Ocean coincident with El Nino.~ J. Geophys. Res. 104, 3035e3047. http://dx.doi. Meltzner, A.J., Sieh, K., Chiang, H.-W., Shen, C.-C., Suwargadi, B.W., Natawidjaja, D.H., org/10.1029/1998JC900085. Philibosian, B., Briggs, R.W., 2012. Persistent termini of 2004- and 2005-like Freymueller, J.T., Fu, Y., Li, S., Ohta, Y., 2014. Geodetic estimates of slow and transient ruptures of the Sunda megathrust. J. Geophys. Res. 117, B04405. http://dx.doi. slip in the Alaska subduction zone. Seismol. Res. Lett. 85, 417. http://dx.doi.org/ org/10.1029/2011JB008888. 10.1785/0220140014. Meltzner, A.J., Sieh, K., Chiang, H.-W., Shen, C.-C., Suwargadi, B.W., Natawidjaja, D.H., Fu, Y., Freymueller, J.T., 2013. Repeated large slow slip events at the southcentral Philibosian, B.E., Briggs, R.W., Galetzka, J., 2010. Coral evidence for earthquake Alaska subduction zone. Earth Planet. Sci. Lett. 375, 303e311. http://dx.doi.org/ recurrence and an A.D. 1390e1455 cluster at the south end of the 2004 10.1016/j.epsl.2013.05.049. AceheAndaman rupture. J. Geophys. Res. 115, B10402. http://dx.doi.org/10. Fu, Y., Freymueller, J.T., Argus, D.F., Owen, S.E., 2014. Temporal variation of a large 1029/2010JB007499. slow slip event at the southcentral Alaska subduction zone during 2009e2013. Natawidjaja, D.H., Sieh, K., Ward, S.N., Cheng, H., Edwards, R.L., Galetzka, J., Seismol. Res. Lett. 85, 444. http://dx.doi.org/10.1785/0220140014. Suwargadi, B.W., 2004. Paleogeodetic records of seismic and aseismic subduc- Gomberg, J., The Cascadia 2007 and Beyond Working Group, 2010. Slow-slip phe- tion from central Sumatran microatolls, Indonesia. J. Geophys. Res. 109, B04306. nomena in Cascadia from 2007 and beyond: a review. Geol. Soc. Am. Bull. 122, http://dx.doi.org/10.1029/2003JB002398. 963e978. http://dx.doi.org/10.1130/B30287.1. Natawidjaja, D.H., Sieh, K., Chlieh, M., Galetzka, J., Suwargadi, B.W., Cheng, H., Great Britain Hydrographic Department, 1934. Malacca Strait Pilot, Comprising Edwards, R.L., Avouac, J.-P., Ward, S.N., 2006. Source parameters of the great Malacca Strait and its Northern Approaches, Singapore Strait and the West Sumatran megathrust earthquakes of 1797 and 1833 inferred from coral Coast of Sumatra. Hydrographic Dept., Admiralty, London, 430 pp. microatolls. J. Geophys. Res. 111, B06403. http://dx.doi.org/10.1029/ Hamlington, B.D., Leben, R.R., Nerem, R.S., Han, W., Kim, K.-Y., 2011. Reconstructing 2005JB004025. sea level using cyclostationary empirical orthogonal functions. J. Geophys. Res. Newcomb, K.R., McCann, W.R., 1987. Seismic history and seismotectonics of the 116, C12015. http://dx.doi.org/10.1029/2011JC007529. Sunda Arc. J. Geophys. Res. 92, 421e439. http://dx.doi.org/10.1029/ Hayes, G.P., Wald, D.J., Johnson, R.L., 2012. Slab1.0: a three-dimensional model of JB092iB01p00421. global subduction zone geometries. J. Geophys. Res. 117, B01302. http://dx.doi. Nishimura, T., Hirasawa, T., Miyazaki, S., Sagiya, T., Tada, T., Miura, S., Tanaka, K., org/10.1029/2011JB008524. 2004. Temporal change of interplate coupling in northeastern Japan during Hill, E.M., Borrero, J.C., Huang, Z., Qiu, Q., Banerjee, P., Natawidjaja, D.H., Elosegui, P., 1995e2002 estimated from continuous GPS observations. Geophys. J. Int. 157, Fritz, H.M., Suwargadi, B.W., Pranantyo, I.R., Li, L., Macpherson, K.A., Skanavis, V., 901e916. http://dx.doi.org/10.1111/j.1365-246X.2004.02159.x. Synolakis, C.E., Sieh, K., 2012. The 2010 MW 7.8 Mentawai earthquake: very Ochi, T., Kato, T., 2013. Depth extent of the long-term slow slip event in the Tokai shallow source of a rare tsunami earthquake determined from tsunami field district, central Japan: a new insight. J. Geophys. Res. 118, 4847e4860. http://dx. survey and near-field GPS data. J. Geophys. Res. 117, B06402. http://dx.doi.org/ doi.org/10.1002/jgrb.50355. 10.1029/2012JB009159. Okada, Y., 1985. Surface deformation due to shear and tensile faults in a half-space. Hsu, Y.-J., Simons, M., Avouac, J.-P., Galetzka, J., Sieh, K., Chlieh, M., Natawidjaja, D., Bull. Seismol. Soc. Am. 75, 1135e1154. Prawirodirdjo, L., Bock, Y., 2006. Frictional afterslip following the 2005 Okada, Y., 1992. Internal deformation due to shear and tensile faults in a half-space. NiaseSimeulue earthquake, Sumatra. Science 312, 1921e1926. http://dx.doi. Bull. Seismol. Soc. Am. 82, 1018e1040. org/10.1126/science.1126960. Ozawa, S., Nishimura, T., Munekane, H., Suito, H., Kobayashi, T., Tobita, M., Hu, Y., Wang, K., 2012. Spherical-Earth finite element model of short-term post- Imakiire, T., 2012. Preceding, coseismic, and postseismic slips of the 2011 seismic deformation following the 2004 Sumatra earthquake. J. Geophys. Res. Tohoku earthquake, Japan. J. Geophys. Res. 117, B07404. http://dx.doi.org/10. 117, B05404. http://dx.doi.org/10.1029/2012JB009153. 1029/2011JB009120. A.J. Meltzner et al. / Quaternary Science Reviews 122 (2015) 258e281 281

Peng, Z., Gomberg, J., 2010. An integrated perspective of the continuum between Sieh, K., Natawidjaja, D.H., Meltzner, A.J., Shen, C.-C., Cheng, H., Li, K.-S., earthquakes and slow-slip phenomena. Nat. Geosci. 3, 599e607. http://dx.doi. Suwargadi, B.W., Galetzka, J., Philibosian, B., Edwards, R.L., 2008. Earthquake org/10.1038/ngeo940. supercycles inferred from sea-level changes recorded in the corals of West Perfettini, H., Avouac, J.-P., Ruegg, J.-C., 2005. Geodetic displacements and after- Sumatra. Science 322, 1674e1678. http://dx.doi.org/10.1126/science.1163589. shocks following the 2001 MW ¼ 8.4 Peru earthquake: implications for the Sieh, K., Philibosian, B., Avouac, J.-P., Natawidjaja, D.H., Chiang, H.-W., Wu, C.-C., mechanics of the earthquake cycle along subduction zones. J. Geophys. Res. 110, Perfettini, H., Shen, C.-C., Daryono, M.R., Suwargadi, B.W., 15e19 December 2014. B09404. http://dx.doi.org/10.1029/2004JB003522. Characterizing the variability of supercycles on the Mentawai segment of the Perfettini, H., Avouac, J.-P., Tavera, H., Kositsky, A., Nocquet, J.-M., Bondoux, F., Sunda megathrust and implications for global fault behavior. Abstract S44C-01 Chlieh, M., Sladen, A., Audin, L., Farber, D.L., Soler, P., 2010. Seismic and aseismic presented at 2014 Fall Meeting. American Geophysical Union, San Francisco, CA. slip on the Central Peru megathrust. Nature 465, 78e81. http://dx.doi.org/10. Sun, T., Wang, K., Iinuma, T., Hino, R., He, J., Fujimoto, H., Kido, M., Osada, Y., 1038/nature09062. Miura, S., Ohta, Y., Hu, Y., 2014. Prevalence of viscoelastic relaxation after the Philibosian, B., Sieh, K., Avouac, J.-P., Natawidjaja, D.H., Chiang, H.-W., Wu, C.-C., 2011 Tohoku-oki earthquake. Nature 514, 84e87. http://dx.doi.org/10.1038/ Perfettini, H., Shen, C.-C., Daryono, M.R., Suwargadi, B.W., 2014. Rupture and nature13778. variable coupling behavior of the Mentawai segment of the Sunda megathrust Taylor, F.W., Frohlich, C., Lecolle, J., Strecker, M., 1987. Analysis of partially emerged during the supercycle culmination of 1797 to 1833. J. Geophys. Res. 119, corals and reef terraces in the central Vanuatu Arc: comparison of contempo- 7258e7287. http://dx.doi.org/10.1002/2014JB011200. rary coseismic and nonseismic with Quaternary vertical movements. Pollitz, F., Banerjee, P., Grijalva, K., Nagarajan, B., Bürgmann, R., 2008. Effect of 3-D J. Geophys. Res. 92, 4905e4933. http://dx.doi.org/10.1029/JB092iB06p04905. viscoelastic structure on post-seismic relaxation from the 2004 M ¼ 9.2 The Singapore Free Press, 1861a, 11 April, p. 3. Sumatra earthquake. Geophys. J. Int. 173, 189e204. http://dx.doi.org/10.1111/j. The Singapore Free Press, 1861b, 25 April, p. 3. 1365-246X.2007.03666.x. The Sydney Morning Herald, 1907, 24 January, p. 7. Prawirodirdjo, L., McCaffrey, R., Chadwell, C.D., Bock, Y., Subarya, C., 2010. Geodetic Tilmann, F.J., Craig, T.J., Grevemeyer, I., Suwargadi, B., Kopp, H., Flueh, E., 2011. The observations of an earthquake cycle at the Sumatra subduction zone: role of updip seismic/aseismic transition of the Sumatra megathrust illuminated by interseismic strain segmentation. J. Geophys. Res. 115, B03414. http://dx.doi. aftershocks of the 2004 AceheAndaman and 2005 Nias events. Geophys. J. Int. org/10.1029/2008JB006139. 187, 539e542. http://dx.doi.org/10.1111/j.1365-246X.2011.05182.x. Rockwell, T., Barka, A., Dawson, T., Akyuz, S., Thorup, K., 2001. Paleoseismology of Uchida, N., Matsuzawa, T., 2013. Pre- and postseismic slow slip surrounding the the Gazikoy-Saros segment of the North Anatolia fault, northwestern Turkey: 2011 Tohoku-oki earthquake rupture. Earth Planet. Sci. Lett. 374, 81e91. http:// comparison of the historical and paleoseismic records, implications of regional dx.doi.org/10.1016/j.epsl.2013.05.021. seismic hazard, and models of earthquake recurrence. J. Seismol. 5, 433e448. van Woesik, R., 2004. Comment on “Coral reef death during the 1997 Indian Ocean http://dx.doi.org/10.1023/A:1011435927983. Dipole linked to Indonesian wildfires”. Science 303, 1297. http://dx.doi.org/10. Savage, J.C., 1983. A dislocation model of strain accumulation and release at a 1126/science.1091983. subduction zone. J. Geophys. Res. 88, 4984e4996. http://dx.doi.org/10.1029/ Wang, K., Hu, Y., He, J., 2012. Deformation cycles of subduction earthquakes in a JB088iB06p04984. viscoelastic Earth. Nature 484, 327e332. http://dx.doi.org/10.1038/nature11032. Savage, J.C., Thatcher, W., 1992. Interseismic deformation at the Nankai Trough, Webster, P.J., Moore, A.M., Loschnigg, J.P., Leben, R.R., 1999. Coupled oceaneatmo- Japan, subduction zone. J. Geophys. Res. 97, 11117e11135. http://dx.doi.org/10. sphere dynamics in the Indian Ocean during 1997e98. Nature 401, 356e360. 1029/92JB00810. http://dx.doi.org/10.1038/43848. Sawai, Y., Satake, K., Kamataki, T., Nasu, H., Shishikura, M., Atwater, B.F., Horton, B.P., Wichmann, A., 1918. Die Erdbeben des Indischen Archipels bis zum Jahre 1857, Kelsey, H.M., Nagumo, T., Yamaguchi, M., 2004. Transient uplift after a 17th- Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, century earthquake along the Kuril subduction zone. Science 306, 1918e1920. Tweede Sectie, deel 20, no. 4. Johannes Müller, Amsterdam, 193 pp. http://dx.doi.org/10.1126/science.1104895. Wichmann, A., 1922. Die Erdbeben des Indischen Archipels von 1858 bis 1877, Scoffin, T.P., Stoddart, D.R., 1978. The nature and significance of microatolls. Philos. Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, Trans. R. Soc. Lond. Ser. B 284, 99e122. http://dx.doi.org/10.1098/rstb.1978. Tweede Sectie, deel 22, no. 5. Koninklijke Akademie van Wetenschappen, 0055. Amsterdam, 209 pp. Shearer, P., Bürgmann, R., 2010. Lessons learned from the 2004 SumatraeAndaman Yokota, Y., Koketsu, K., 2015. A very long-term transient event preceding the 2011 megathrust rupture. Annu. Rev. Earth Planet. Sci. 38, 103e131. http://dx.doi.org/ Tohoku earthquake. Nat. Commun. 6, 5934. http://dx.doi.org/10.1038/ 10.1146/annurev-earth-040809-152537. ncomms6934. Shen, C.-C., Cheng, H., Edwards, R.L., Moran, S.B., Edmonds, H.N., Hoff, J.A., Yu, L., Rienecker, M.M., 1999. Mechanisms for the Indian Ocean warming during the Thomas, R.B., 2003. Measurement of attogram quantities of 231Pa in dissolved 1997e98 El Nino.~ Geophys. Res. Lett. 26, 735e738. http://dx.doi.org/10.1029/ and particulate fractions of seawater by isotope dilution thermal ionization 1999GL900072. mass spectroscopy. Anal. Chem. 75, 1075e1079. http://dx.doi.org/10.1021/ Zachariasen, J., Sieh, K., Taylor, F.W., Hantoro, W.S., 2000. Modern vertical defor- ac026247r. mation above the Sumatran subduction zone: paleogeodetic insights from coral Shen, C.-C., Edwards, R.L., Cheng, H., Dorale, J.A., Thomas, R.B., Moran, S.B., microatolls. Bull. Seismol. Soc. Am. 90, 897e913. http://dx.doi.org/10.1785/ Weinstein, S.E., Edmonds, H.N., 2002. Uranium and thorium isotopic and con- 0119980016. centration measurements by magnetic sector inductively coupled plasma mass Zurcher, F., Margolle, E., 1866. Volcans et Tremblements de Terre. Hachette et Cie, spectrometry. Chem. Geol. 185, 165e178. http://dx.doi.org/10.1016/S0009- Paris, 372 pp. Republished by Adamant Media Corporation, 2006. Newer edition 2541(01)00404-1. available at: http://books.google.com/books?id¼iQcKAAAAIAAJ&pg¼PP9 Shen, C.-C., Wu, C.-C., Cheng, H., Edwards, R.L., Hsieh, Y.-T., Gallet, S., Chang, C.-C., https://archive.org/details/volcansettrembl01marggoog. Li, T.-Y., Lam, D.D., Kano, A., Hori, M., Spotl,€ C., 2012. High-precision and high- Zurcher, F., Margolle, E., 1868. Volcanoes and Earthquakes. Translated by Lockyer, W. resolution carbonate 230Th dating by MC-ICP-MS with SEM protocols. Geo- Richard Bentley, London, 261 pp. Republished by Cambridge University Press, chim. Cosmochim. Acta 99, 71e86. http://dx.doi.org/10.1016/j.gca.2012.09.018. New York, 2012. Available at: http://dx.doi.org/10.1017/CBO9781139226806 Shen, C.-C., Li, K.-S., Sieh, K., Natawidjaja, D., Cheng, H., Wang, X., Edwards, R.L., https://archive.org/details/volcanoesearthqu00zurcrich. Lam, D.D., Hsieh, Y.-T., Fan, T.-Y., Meltzner, A.J., Taylor, F.W., Quinn, T.M., Zweck, C., Freymueller, J.T., Cohen, S.C., 2002. The 1964 great Alaska earthquake: Chiang, H.-W., Kilbourne, K.H., 2008. Variation of initial 230Th/232Th and limits present day and cumulative postseismic deformation in the western Kenai of high precision U-Th dating of shallow-water corals. Geochim. Cosmochim. Peninsula. Phys. Earth Planet. Inter 132, 5e20. http://dx.doi.org/10.1016/S0031- Acta 72, 4201e4223. http://dx.doi.org/10.1016/j.gca.2008.06.011. 9201(02)00041-9.