3.1 IBD in Java
We discovered seven supratidal IBD sites in Java (Fig. 2, Table 1). There are likely others either covered in sand or in parts of the coast we could not access. The most significant discovery was at Papuma (site J6, Fig. 1), a beach impacted by a tsunami from the 1994 7.9 Mw earthquake in east Java (Fig. 1). Although no tsunami height measurements were made at Papuma, nearby beaches recorded maximum tsunami heights from 11–14 m (Synolakis et al. 1995; Tsuji et al. 1995). By the time of our survey, sand deposits from the tsunami were scarce. However, residents and fishermen claim that the IBD at Papuma (Fig. 4a), and another site to the east at Pasir Putih (site J7), formed during the tsunami. Although the IBD were not mentioned during post-tsunami surveys (Maramai and Tinti 1997; Synolakis et al. 1995; Tsuji et al. 1995b) residents who witnessed the tsunami independently provided first-hand accounts of changes to the coastline that occurred including the formation of the IBD.
Papuma beach consists of outcrops of beachrock in the intertidal zone with layers that differ in resistance to erosion (Fig. 4). The outcrops show evidence of plucking of beachrock slabs from the intertidal zone, with pluck marks that match the size and shape of nearby detached slabs (Fig. 4). The minimum distance between sites of beachrock excavation and the IBD is around 25 meters up a beach slope of 5 degrees. At the base of the IBD the slope steepens abruptly to as much as 10–18 degrees (Fig. 4). Individual boulders in the Papuma IBD have consistent average strike directions within 6 degrees of the strike of the shoreline (Meservy (2017). After the tsunami, residents observed that some rock slabs were found on the beach road and further inland of the platform backstop.
Table 1
Location and characteristics of IBD.
Place/Beach
|
Site
|
Latitude
|
Longitude
|
a-axis
m
|
b-axis
m
|
c-axis
m
|
aFb
|
Area
m2
|
Elev
m
|
Java
|
|
|
|
|
|
|
|
|
|
Nampu
|
J1
|
-8.210442
|
110.904169
|
2
|
1.2
|
0.3
|
5.3
|
2.4
|
3
|
Klayar
|
J2
|
-8.223617
|
110.947155
|
3.3
|
1.7
|
1.8
|
1.4
|
5.8
|
4
|
Blosok
|
J3
|
-8.234227
|
110.965912
|
3.4
|
2.2
|
0.3
|
9.3
|
9.6
|
4
|
Pedenombo
|
J4
|
-8.241134
|
110.983327
|
1.8
|
1.4
|
0.3
|
5.3
|
2.4
|
5
|
Pidakan
|
J5
|
-8.255659
|
111.238847
|
2.1
|
1.4
|
0.2
|
8.8
|
3.3
|
2
|
Papuma
|
J6
|
-8.435089
|
113.551749
|
2.4
|
1.9
|
0.3
|
7.2
|
4.7
|
4
|
Pasir Putih
|
J7
|
-8.561527
|
113.924558
|
0.8
|
0.7
|
0.5
|
1.5
|
0.6
|
|
Bali
|
|
|
|
|
|
|
|
|
|
Pandawa
|
B1
|
-8.843602
|
115.189514
|
2.6
|
1.8
|
0.3
|
7.3
|
4.6
|
3
|
Lombok
|
|
|
|
|
|
|
|
|
|
Are Guling
|
L1
|
-8.913288
|
116.243961
|
2.5
|
1.3
|
0.3
|
6.3
|
3.2
|
4
|
Putinyale
|
L3
|
-8.909685
|
116.298494
|
2.2
|
1.8
|
0.4
|
5.0
|
4
|
4
|
Payung
|
L4
|
-8.917953
|
116.329526
|
3.6
|
2.8
|
0.6
|
5.3
|
10.3
|
4
|
Kura-Kura
|
L5
|
-8.919369
|
116.441725
|
3.4
|
2.3
|
0.3
|
9.5
|
7.9
|
5
|
Sumba
|
|
|
|
|
|
|
|
|
|
Maloba
|
S1
|
-9.777156
|
119.651881
|
2.5
|
1.8
|
0.5
|
4.3
|
3.9
|
5
|
Kisar
|
|
|
|
|
|
|
|
|
|
SW coast
|
K1
|
-8.050024
|
127.138544
|
1.8
|
1.2
|
0.6
|
2.5
|
2.2
|
8
|
SW coast
|
K2
|
-8.089577
|
127.146065
|
|
|
|
|
|
20
|
SW coast
|
K3
|
-8.097237
|
127.145059
|
|
|
|
|
|
19
|
Leti
|
Lt1
|
-8.213141
|
127.60159
|
3.2
|
1.6
|
0.6
|
4.0
|
5.1
|
4
|
Nailaka
|
Bn1
|
-4.535583
|
129.696464
|
2.7
|
1.5
|
0.4
|
5.3
|
4.1
|
3
|
a – Flatness Index (Fb) = (a + b) / 2c
Site J7 (Pasir Putih), which is in a nationally protected forest, translates to “White Sand Beach”. However, the 1994 tsunami removed all the sand from the beach and transformed it into one entirely covered by small (< 0.6 m2), subrounded imbricated boulders and cobbles of that spill several meters over the platform backstop into the forest. The tsunami deposited what boulders were available offshore, which did not include exposed beachrock. This pattern of no IBD where there is little to no beachrock is common if not ubiquitous. Further west along the southern Java coast, in the Pacitan region, we discovered five additional IBD sites (Fig. 2, 4). Most of the beaches we investigated have a wide, flat platform that extends 150–200 meters seaward from the toe of the beach to the edge of barrier reef. The platforms are easily traversed during low tide because the large waves at these locations break offshore against the fringing reef. Large blocks of broken coral litter the outer edge of the platform, some of which have moved shoreward (Fig. 5).
According to Scheffers (2008) a shallow water platform setting makes it difficult for wind waves to move boulders and rock slabs due to loss of energy not only when the waves break at the fringing reef edge, but also as they travel across the width of the shallow platform. The boulders at each site, and source areas for the boulders, are mostly beachrock like what we observe at Papuma (Fig. 4). Some beaches have slabs of limestone from nearby bedrock outcrops and coral mixed in with the IBD. Well exposed areas of beachrock excavation show ramp-flat excavation plane geometries like those in imbricate thrust sheets (Fig. 3, 5a). Some slabs of beachrock have left scour marks in their wake that are traceable for > 100 m across the intertidal platform before deposition (Fig. 5b). Most exposed areas of beachrock in the intertidal zone are adjacent to IBD.
3.2 IBD East of Java
East of Java, IBD are found along the southern coastlines of Bali, Lombok, and Sumba of the eastern Sunda Arc, and along coastlines of the Banda Arc Islands of Kisar, Leti and Nailaka (Fig. 1). The Sunda Arc Islands face the Java Trench while the Banda Arc Islands face the Timor and Tanimbar Troughs, which is influenced by continental subduction (e.g., Harris, 2011). In Lombok IBD are up to 750 m long and 25 m wide at Kura-Kura Beach (site L4) in SE Lombok (Fig. 1, 4). Similar characteristics are found in these deposits as documented from the known tsunami deposit in Papuma.
3.3 Composition of IBD
The IBD we discovered mostly consist of slabs of dense, carbonate cemented sand and gravel of both clastic and biogenic origin (Fig. 7), which is characteristic of beachrock in general (e.g., Russell 1963; Bricker 1971; Milliman 1974; Vieira et al. 2007). Most beachrock is documented in mid-low latitude coastal regions and beaches with low tidal variation (Vousdoukas et al. 2007). Cementation of the beachrock mostly happens in the vadose zone during low tide when the sand deposits are exposed to the atmosphere. The lithification process is related to cementation of carbonate from fresh water (calcite) or sea water (aragonite) in the intertidal zone. Some studies show a strong correlation between beachrock formation and proximity to carbonate-rich shorelines. Most of the IBD sites in the eastern Sunda Arc are near carbonate bedrock. Although beachrock is a common feature of coastlines globally, ridges of stacked or imbricated beachrock are rare except in tectonically active areas that are likely influenced by recurring tsunami.
Compositional exceptions to IBD are accumulations of boulders sourced from wave-cut cliffs. These varying-shaped boulders are also imbricated by high energy events (Fig. 7). IBD lacking beachrock are found at Nampu (site J1) and Klayar (site J2) in Java and Payung (L1) in Lombok (Fig. 2).
3.4 Size and orientation of IBD
Evidence of rapid undercutting and recent bedrock cliff failure and retreat is visible throughout much of southern coastlines of Java, Bali, Lombok, and the lesser Sunda Islands. Typically, when a coastal cliff collapses, or a landslide occurs near the coast, rocks from onshore are initially distributed randomly along the coastal bench. Large waves can break up some of these boulders and stack them in with beachrock slabs (Fig. 7).
In contrast, beachrock from the intertidal zone of the seafloor stacks up in a different way to produce IBD. In situ beach rock is constantly being attacked by storm waves causing differential erosion of some layers that introduce instabilities, such as overhangs that cause beachrock to break into allochthonous slabs (Fig. 6). Large waves transport most slabs from their source in the intertidal zone to near the maximum high tide mark to form IBD. Nearly all the stacked slabs of IBD are platy with average long a-axes of 2.8 m, intermediate axes (b-axis) of 1.8 m and thicknesses (c-axis) of 40 cm. Volumes of the largest boulders are 3–4 cubic m (Table 1). In general, the largest rock slabs are around the same size, with exceptions found at sites J4 and L4 (Table 1).
Most IBD have a preferred strike orientation of the a-b plane, which is sub-parallel to the strike of the shoreline within 2–10 degrees (Meservy 2017). All the beach rock slab accumulations have similar, right-skewed distributions with between 0.5-1.0 cubic meters in volume. Of the largest boulders, none are much greater than 3 cubic meters. Whether this size limit is a function of wave height or speed, or from the size of the boulders available is unknown (Benner et al 2010). Other studies demonstrate how boulder sizes do not correlate well with known wave speeds that deposited them (Etienne et al., 2011). More important are the sizes of the boulders available for the wave to entrain (Scheffers 2021). Nott (2003) notes that during the 1998 Sissano tsunami in Papua New Guinea a concrete slab with a- b- and c-axes of 3.15, 1.22 and 0.23, respectively, moved 400 meters inland from flow depths of 5 m. This depth is less than those inferred for Papuma.
3.5 Uncommon storm wave height experiment
IBD provide conditions for conducting experiments to test the contribution of storm waves in forming or shaping IBD (Oak 1984, Lorang 2002 Goto et al. 2007, Goto et al. 2009b, Benner et al. 2010; Goto et al. 2010a, Goto et al. 2010b, Etienne and Paris 2010, Goto et al. 2011; Schneider et al. 2019). Between 1983 and 2016 there have been 5 high wave events with an average of 4.7 m swells and periods of 18 − 10 seconds (Surfline.com) along the southern coastlines of eastern Sunda Arc islands.
To measure possible boulder movement during these storms we constructed UAV-assisted digital surface models (DSMs) with a high degree of boulder registration through time at Pedenombo beach (Fig. 8). This beach consists of 3 platforms. The lowest is the current wave cut platform forming the intertidal zone. The second is 3 m above sea level and likely represents a wind wave terrace. This platform hosts the IBD, which stack up against the shoreward edge of platform 3. The third platform could be an uplifted Holocene coral terrace or a slightly subsided 5e sea level high-stand terrace formed at approximately 120 ka.
High resolution UAV data were collected during four separate missions between 2016 and 2019. The first pre- and post-storm measurements were taken on 7/30–31/2016 and 8/2/2016, which was immediately before and after a 4.2 m swell that struck the beach during a + 2.5 m spring high tide (Surfline.com). The third set of drone images were taken on 7/12/2017 after an epoch of 342 days. The fourth set was taken 6/20/2019 after an epoch of 708 days. During this nearly two year period the beach was impacted by waves from two offshore tropical cyclones.
3.6 Discussion of DSM Overlays at Pedenombo
Image (a) is an orthophoto documenting the initial positions of IBD slabs 1 day before an uncommon storm wave event (Fig. 9a). Additionally, 21 limestone boulders < 0.5 m in diameter were marked, manually placed, and surveyed seaward of the IBD. See small open circles in Fig. 9b for original positions of hand placed boulders.
Image (b) is an overlay of DSMs from before and after the 6.7 m high wave event on 7/30–31/2016 (Fig. 9b). The image documents the initial and final positions of manually placed boulders and pre-existing beachrock slabs of the IBD (Fig. 11b). After the storm, we located all but two of the boulders and resurveyed their positions. Green circles show pre- and red triangles show post-storm locations. Lines on the image connect initial and final boulder positions and reveal that all manually placed boulders moved, if only 10 cm. Most boulders moved laterally along the shoreline, even off the image from 2–45 m. Five placed boulders were added to the front of the IBD, but none moved onto the IBD. At least 8 pre-existing beachrock slabs of the IBD moved < 2 m as indicated by the proximity of slab excavation and addition marks. The mirror image of some of these marks indicate slabs flipping about their a-axes.
Image (c) is an overlay of DSMs of the post-storm 2016 and 2017 UAV surveys. During this time waves from 2.2 m to 2.6 m were recorded on June 23, 2017 (Avia 2020). At least 19 IBD slabs moved < 2 m based on matching of green and nearby red shapes (Fig. 9c). The mirror image of some of these shapes indicate slab flipping about the a-axis as in Fig. 9b. At least 8 other slabs circled in green have ambiguous sources. These slabs may have been added to the IBD by storm waves (not likely based on other observations) or their original position is eroded or filled in by storm deposited sand, gravel, and cobbles. Changes in sand volume is detected at ten localized zones on the beach with an average volume change of approximately 65 m2.
Image (d) is an overlay of DSM (c) and (d), which records the effects of some of the most energetic waves over the most time (two offshore tropical cyclones). During this time, overlapping category 1 cyclones Cempaka and Dahlia brushed by Java increasing wave heights from 1.8 m to 3.2 m (Windupranata et al. 2019; Avia 2020). Cempaka, which nearly made landfall in the Pedenombo beach area, affected 13 coastal communities in the Pacitan area. In total there were 41 deaths, 20,000 people evacuated, and around US$83.6 million in damages (Badan Nasional Penanggulangan Bencana). Although the final overlay is messy due to clastic sediment redistribution, it reveals that around 113 individual beachrock slabs moved slightly or flipped of the approximately 1220 slabs in the IBD. One boulder in the lower left corner of image (c) disappears in image (d). However, there is no conclusive evidence of boulder addition or removal. UAV-based post-tsunami surveys of the 2004 Indian Ocean Tsunami also found that no boulders were added to or removed from the tsunami formed IBD by monsoonal storms (Etienne et al. 2011). These results provide limits for Indonesian storm waves to move boulders and strongly support an interpretation that the IBD are primarily formed by tsunamis. However, it is important to recognize the role storm waves play in eroding beachrock in the intertidal zone into detached slabs that can be transported later by tsunami waves.
3.7 Radiocarbon analyses
Another way to test the relative contribution of storm versus tsunami waves in forming IBD is age analysis of coral interlayered with the beachrock slabs. These ages likely differ depending on whether coral boulders were emplaced by storms or tsunamis. Tsunami ages may cluster around the time of known historical or paleo-tsunami event whereas storm wave deposits could have a much broader age distribution since coral boulders could be added yearly. Dating IBD is problematic due to the likelihood of mixing offshore siliciclastic and bioclastic material of various ages (e.g., Mastronuzzi et al. 2000; Ishizawa et al. 2020).
Our reconnaissance age analysis of several coral boulders yielded reliable ages at 5 sites (Table 2). Although there is some scatter, most ages correspond closely with known historical tsunamis and candidate paleo-tsunamis in the eastern Sunda Arc (Sulaeman, 2018). Only six tsunamis affecting eastern Sunda Arc islands are recorded in historical accounts dating back to 1584 (Harris and Major 2016). Lack of possible events reduces the likelihood of coral boulder ages coinciding with known and inferred tsunamis by chance (Table 2). For example, the 1861 AD age of coral boulders in IBD near Pacitan is concordant with the 1859 Pacitan tsunami (Harris and Major 2016). The 1692 age of another boulder coincides with one of the largest earthquakes to strike Java documented in historical accounts. This event was felt widely throughout the region with a maximum MMI of X, had at least 13 months of aftershocks, and a tsunami of unknown extent (Harris and Major 2016). The 1474 AD age correlates with a 400–600 years BP OSL age of a candidate tsunami sand deposit at 1 meter depth in Bali (Sulaeman 2018). Similar high-energy sands with marine debris occur onshore at around the same depth throughout the southern coastal plains of the eastern Sunda Arc.
The wide spectrum of ages for the imbricate coral boulders we sampled across the five, widely spread beaches may result from many factors, which is not uncommon for coastal boulder accumulations (e.g., Vousdoukas et al., 2007). The age ranges may indicate that the IBDs have amassed over a series of events (Nott 1997 and 2004). Notwithstanding these issues, we point out that it is highly unlikely that most boulder ages are concordant with the limited number of events reported in historical records and the two proposed paleo-tsunami events in region. In fact, only two ages that do not correlate are those of 500–700 AD (Table 2). If these boulder ages correspond to a tsunami they are consistent with a ~ 500 year recurrence interval of mega-thrust earthquakes.
Table 2
Radiocarbon chronology of shells and coral boulders found wedged within five IBD in Java.
Location
|
Sample
Type
|
Calibrate
Calendar Age
|
95% Confidence
|
Possible Event
|
Binuangeun
West Java
|
shell
|
1053 AD
|
990–1153 AD
|
1200 − 800 AD
Java Trench**
|
|
shell
|
991 AD
|
908–1050 AD
|
1200 − 800 AD
Java Trench**
|
|
coral boulder
|
989 AD
|
907–1049 AD
|
1200 − 800 AD Java Trench**
|
|
shell
|
905 AD
|
806–994 AD
|
1200 − 800 AD
Java Trench**
|
|
Coral boulder
|
663 AD
|
596–724 AD
|
?
|
Blosok-C
Site J3
|
coral boulder
|
1861 AD
|
1810–1880 AD
|
1859 AD Pacitan
|
Blosok-B
Site J3
|
coral boulder
|
1692 AD
|
1640–1720 AD
|
1699 AD Java
|
Blosok-A
Site J3
|
coral boulder
|
551 AD
|
500–580 AD
|
?
|
Pidakan
Site J5
|
coral boulder
|
1474 AD
|
1440–1510 AD
|
1400–1600 AD
Java Trench**
|
Pedenombo
Site J4
|
Shell on boulder
|
1177 AD
|
1080–1220 AD
|
1200 − 800 AD
Java Trench**
|
Kisar 1*
Site K1
|
coral boulder
|
1903 AD
|
1898–1908 AD
|
1896 AD Timor
|
Kisar 2*
Site K2
|
coral boulder
|
1904 AD
|
1897–1909 AD
|
1896 AD Timor
|
* Boulder deposit on coral terrace, Major et al. (2013) ** Sulaeman (2018) *** |
The discovery of several tsunami related IBD along a 1200 km section of the south coastal regions of the eastern Sunda Arc (Sulaeman 2018) that are likely emplaced by tsunamis indicates that this region is as susceptible to recurring mega-thrust earthquakes and large tsunamis as the western Sunda Arc (Sumatra). Candidate paleo-tsunami sand deposits discovered throughout the eastern Sunda Arc also attest to at least two large, multi-island tsunamis at around 1000 and 1500 AD (Sulaeman 2018). These findings are consistent with ages for some of the shells and coral boulders incorporated into the IBD of the Sunda Arc and NW Australia. We also found relatively young boulders deposited up to 20 meters above sea level on the 125 ka uplifted coral terrace in Kisar (Major et al. 2013). These ages are concordant with the 1896 earthquake and tsunami in Timor, which is the only major earthquake from the Timor region found in over 400 years of historical records (Harris and Major 2016).
3.8 NW Australia IBD from Java Trench Earthquakes?
IBD and large boulders at elevations up to 20 m along the 2500 km of coastline of NW Australia document huge waves impacting the coast before historical times (Nott and Bryant 2003; Goff and Chague-Goff 2014). Imbricated boulder ridges, like those we discovered along the southern coasts of the eastern Sunda Arc islands, are documented up to 6 m elevation. The largest boulder measured at the Australia IBD has an a-axis of 5 m, but the average boulder sizes are a = 3.0, b = 2.1, c = 0.7 m (Nott, 2004). Shells at the back of the ridge have radiocarbon chronologies dating around the mid to late 19th century, a period of heavy seismicity near Pacitan (Harris and Major 2016). Other ages roughly align with those of pre-historic tsunamis (Table 2).
Tsunamis that have impacted NW Australia during the last century were generated by the 1977 Sumba earthquake (Mw 8.3, Pradjoko et al 2015), and the 1994 (Mw 7.8) and 2006 (Mw 7.7) Java Trench earthquakes (Burbridge et al 2009). Runup heights in NW Australia from the 1977 and 1994 tsunamis were around 4 m (Nott and Bryant 2003). The 1994 tsunami transported coral boulders and marine fauna over 1000 m inland. The 2006 tsunami, which was further north on the Java Trench, had a flow depth of 1–2 m, maximum runup of 8 m, and inundated around 100–200 m inland (Prendergast and Brown 2011). None of the tsunamis formed new or noticeably modified existing IBD.
The NW Australian IBD at Exmouth consist mostly of beachrock like those we discovered in the eastern Sunda Arc. The fact that the 1994 Java Trench tsunami formed an IBD along the Java coast, but not in Australia may indicate that the Australia deposits require larger tsunamis from mega-thrust earthquake events. According to the age analysis of the boulders in the Exmouth IBD, a tsunami of this scale happens over time intervals of 400 to 500 years (Nott 2004), which is consistent with age distributions in the eastern Sunda ARC (Table 2). The age results from Australia and eastern Sunda Arc also argue strongly against IBD being formed by cyclones, which happen on a much more regular basis.
To test the potential of cyclones forming the NW Australia IBD, Nott (2004) conducted pre- and post- cyclone visual surveys. One of the mega-storms was TC Vance, which is the most intense category 5 cyclone ever recorded to cross Australian shores. Surveys were also taken after 3 other category 5 cyclones. These surveys found no evidence of IBD development even in areas where abundant loose boulders were available. Nott (2004) also claims that existing IBD were not disturbed by the mega-storms and no new beachrock slabs were added or removed. It is possible that IBD from the Indonesian locations described in this paper and the ones at Exmouth, Australia may be emplaced by tsunamis from the same earthquakes with those of Australia only recording megathrust events on the Java Trench.
3.8 Tsunami modelling and inundation maps
If IBD in the eastern Sunda Arc and northern Australia record mega-thrust earthquake-caused tsunamis on the Java Trench, most of which have no historic precedent, then tsunami hazards risk assessments should consider the likelihood and impact of these events. To investigate this impact, we constructed tsunami models for two earthquake scenarios along the Java Trench, both of which could potentially form IBD simultaneously along the south coast of the eastern Sunda Arc islands and in NW Australia. We include estimates of the number of people currently inhabiting the areas of inundation in the models.
Model 1 is for an earthquake that fills the seismic gap between the 1994 and 2006 subduction interface earthquakes on the Java trench south of Java (Figs. 1 and 10). This gap could have ruptured to produce the 1859 and 1699 earthquakes and tsunamis noted in the Pacitan area where (Harris and Major 2016). The slip distribution from the 2011 East Japan earthquake is used to model an Mw = 8.4 earthquake on the subduction interface between the 1994 and 2006 tsunamigenic earthquakes. This location is also near where most of the IBD in Java are found (Fig. 12).
Model 2 (Fig. 11) is a worst-case scenario Mw = 9.0 mega-thrust earthquake that ruptures the seismic gap along the entire Java Trench. The amount of slip in this model is based on the sudden release of around 28 m of elastic strain or tectonic loading that has already accumulated on the 1200 m length Java Trench during the past 478 years (Harris and Major 2016). The result is at least a Mw 9.0 mega-thrust earthquake and associated tsunami like the December 2004 or the March 2011 East Japan mega-thrust events. The transoceanic propagation of this event (Fig. 12) shows the highest waves striking the western most tip of Australia where most of the IBD reported from this region are observed Nott (2000).