Reconstructing the geomorphic history of a cave: a case study from Liang Bua, Indonesia part of Vignettes:Vignette Collection
Liang Bua, a large cave situated 16 km from Ruteng in western Flores, Indonesia (Figure 1), was formed as a subterranean chamber over 600 ka. From this time to the present, a series of geomorphic events influenced the structure of the cave and deposits (revealed by excavations SI-SVII, Figure 2), creating a complex stratigraphy (Figure 3) containing stone artifacts, plant and animal remains, pottery, metal items and skeletal remains including Homo floresiensis. Within these deposits, nine main sedimentary units have been identified (Figure 3). The stratigraphic relationships between these units provide the evidence needed to reconstruct the geomorphic history of the cave. This history was dictated by rapid uplift and dominated by water action including slope wash processes, channel formation, pooling of water, and flowstone precipitation, which created waterfalls, cut-and-fill stratigraphy, large pools of water, and extensive flowstone cappings. The sediments in the cave can be divided into two areas: the rear chamber, comprising mainly of in situ and modified conglomerate and slope wash deposits (Figure 2), and the front chamber, consisting of deeply stratified and complex interbedded deposits (Figures 2 and 3). Unit 1 is a densely packed clast-supported conglomerate with a coarse matrix of silty sand (Figure 2). It represents the river's first high-energy entry into the cave and the oldest unit at ~193 ka according to the results of red TL dating techniques. The river deposited conglomerate in the rear and domed front chamber and during the last interglacial (~130-120 ka), the unit was modified by extensive erosion and the deposition of slope wash (Figure 4a). Unit 2 is a basal unit consisting of fine silts and clays and contains lenses of coarse sand, low-angle cross-bedding of coarse and fine units, with some non-parallel and convoluted bedding (Figure 3, Unit 2). This unit was deposited by suspension at ~100 ka in a pond that was derived from slopewash from other caves in the system. Unit 3 is collapse material (fallen pressure slabs) and slumped channel infill containing stone artifacts and bone at relatively high densities (Figure 3, Unit 3). It was deposited in scoured channels and represents the oldest intensive occupation layer. The collapse material from the evolution of the domed cave roof created a dam that filled with channel infill ~95 ka. This age range supports the notion that the river had stopped entering the cave, and the ponding events had ceased by ~100 ka (Figure 4b). Unit 4 is an occupation deposit with high concentrations of stone artifacts and bone, including those of neonatal Stegodon, giant tortoise, and Homo (Figure 2, Unit 4). It contains a matrix of silty clay capped by flowstone and represents the second main layer of occupation. The arrangement of pools and channels during this period almost surrounded an area of cave floor directly to the southwest of Sector IV (Figure 1), which after a period of channel erosion created a greater relief and formed a remnant area of higher ground. This area is thought to have been a zone of occupation, representing a flat, dry haven within the cave environment (Figure 4c). Unit 5 is a fine clayey-silt suspension deposit that accumulated in a pool created by the decline of an originally erosive, wide and deep channel by the east wall of the cave ~55 ka (Figure 2, Unit 5). Unit 6 is a reworked conglomerate eroded from the easterly corner of the rear and front chambers and deposited in the newly formed pool between 55-41 ka (Figure 2, Unit 6). Unit 7 is a clayey-silt occupation deposit containing the LB1 skeleton (holotype of Homo floresiensis) and contains a further phase of intensive occupation. This unit represents a pool deposit containing the skeleton and bank deposits containing stone artifacts, hearth stones, charcoal, and bone, including Stegodon, Komodo dragon, and other hominin remains (Figure 2, Unit 7). The pool by the east wall persisted in the cave environment for at least ~ 40 ka from 55-11 ka, and the absence of sedimentary structures suggests that either the pool varied in depth, or the rear sinkholes and resulting waterfall was a source of intermittent disturbance and sedimentary infilling. The presence of a pool at this time is one of the reasons for the excellent preservation of the skeletal evidence, as once emplaced it was rapidly covered in fine silts during a period of channel inactivity. The bank deposits of the pool also represented an area of higher ground and a zone of occupation ~18 ka (Figure 4d). Unit 8 is a volcanic deposit consisting of a coarse black volcanic sand and fine silt layers containing tephra that both represent primary airfall deposits and were deposited in pools by suspension (Figure 2, Unit 8). Between ~18-16 ka the black sand unit was deposited by the east wall and represents a localized eruption. Between 12-11 ka the pool acted as a sink for accumulating tephra deposits from an extensive volcanic eruption that is more regional in its origin. Unit 9 is a younger occupation and modern slope wash deposit consisting of clayey silts and sandy silts and contains charcoal, fire-reddened clays, Neolithic material including pottery, burials, and a high density of artifacts (Figure 2, Unit 9). It was deposited at ~10 ka when the pools and channels had disappeared, sheetwash replaced rill wash and overland flow, resulting in the horizontal deposition of clayey silts. This process removed the uneven topography and created a level cave-floor surface that was better suited to occupation. The sequence of geomorphological events interpreted from these units (e.g., exposure of cave and high-energy deposition, suspension in a pool, creation of higher ground, channel erosion, cut and fill, bank and pool deposition, volcanic events and sheetwash) restricted occupation to certain zones (Figure 4). Continuous occupation of the cave may not have been possible until after ~100 ka, when the accumulated water had drained from the front chamber. The next ~89 ka were dominated by flowing water creating waterfalls, cut-and-fill by channels, periodic pooling of water, and occasional volcanic events and the creation of zones of occupation - two earlier zones established ~74-61 ka by the west wall and center (Figure 4c) and a later zone ~18 ka and located by the east wall (Figure 4d).
Establishing cave exposure using evidence from river deposits: a case study from Liang Bua, Indonesia part of Vignettes:Vignette Collection
Liang Bua, in western Flores, Indonesia (Figure 1) is a large limestone chamber with a wide entrance and a cathedral-like atmosphere. It's name, meaning 'cool cave' explains its attraction as an occupation site for early hominids. But the cave hasn't always been so appealing. Liang Bua was originally a subterranean chamber and was later exposed by landscape processes, such as river downcutting and cave collapse that were fuelled by rapid tectonic uplift. Presently, the site is situated at the same elevation as the highest of three alluvial terraces (510 m), which were deposited by the Wae Racang river in a wide Miocene limestone valley. The sub-aerial exposure of Liang Bua created the first opportunity for human occupation and since then the cave has accumulated a wealth of archaeological material during a ~100 ka occupation period, including the almost complete skeleton of Homo floresiensis. Establishing when the cave was exposed provides an understanding of how landscape processes interact within a tropical karst valley and helps us to establish a maximum age of occupation for the site. Evidence for the exposure of Liang Bua can be gained from: 1) the age of the adjoining river terraces outside the cave; 2) the pattern of cave collapse; and 3) the age of alluvial deposits found in the cave. 1. The age of the stepped river terraces. The Wae Racang flows for ~25 km from the slopes of Ruteng town north to the sea at Reo (Figure 1b, c). Tectonic uplift in this region has triggered rapid river downcutting and the abandonment of the original floodplain creating a series of three stepped alluvial terraces (Figure 2). The first has been extensively eroded by river action in the vicinity of Liang Bua and has suffered limestone collapse, preserving only a remnant of its original extent. On the northern side of the valley, the third terrace is extensive, with only a dry valley running westwards to interrupt its form. The second terrace is most apparent in front of Liang Bua, where it laterally extends for 62 m along the flanks of the river valley. On the opposite side of the river the second terrace is absent and the first terrace extends straight down to the present river channel. The third alluvial terrace is a Holocene deposit that extends down to the Wae Racang (474 m), but it is absent directly adjacent to Liang Bua. The ages of these terraces, provided by luminescence dating of sediment, combined with key sedimentary characteristics and terrace morphology, provides a timescale for the location of the river during terrace development (Figure 3). The river would have been able to reach the same elevation as the cave, via overbank deposition of the first terrace between 250-215 ka, suggesting that the river occupied the same elevation as the cave ~200 ka. It subsequently eroded its floodplain over the next 100 ka, reaching the elevation of the second terrace at ~18 ka followed by the third terrace during the Holocene at ~5 ka. The presence of a largely eroded first terrace on the southern side of Wae Racang valley, and a well-preserved corresponding terrace on the opposite side of the valley, suggests that the river had wandered across to the southern side of the valley ~200 ka, prior to terrace formation. A layer of pyroclastic deposits capped the third terrace and aided its preservation on the northern side of the valley, while a section of imbricated gravel implies that this location represented a mid-point as the river wandered across to the southern valley. The extensive preservation of the second terrace on the southern side of the river suggests that when the river occupied this elevation at ~20 ka it had returned to the midpoint in the valley. The sedimentological characteristics of these terraces help to reconstruct the pattern of river movement and determine the location of the river during certain time periods. This evidence is used to identify a mechanism for cave exposure. 2. The pattern of cave collapse. The soft, impure, tuff-bearing clastic limestones of this region facilitate high densities of solution features that are prone to collapse. Tectonic uplift encourages valley deepening and weakens the structure of dissolution chambers, which is the primary cause of collapse in this landscape (Figure 4). Evidence of this process can be observed in the sediments of Liang Bua at ~5 m, where the presence of huge limestone blocks suggests that the cave entrance underwent large-scale collapse. This process was important for initiating the accumulation of archaeological evidence as it transformed a discrete opening into a wide entrance, thereby creating an environment suitable for human occupation. 3. The age of the alluvial conglomerate deposits An extensive alluvial conglomerate dominates the southern end of Liang Bua, extending 18 m from the rear wall. The volume of this deposit, its location, and the calibre and composition of its clasts indicate high-energy fluvial transport and deposition of allogenic sediments. The deposition of these fluvial sediments represents the first exposure of the cave by the Wae Racang (Figure 4). The age of the conglomerate cliff exposure ~190 ka indicates that sub-aerial exposure of the cave occurred no earlier than the late Middle Pleistocene, which agrees with the age range of the first alluvial terrace at ~200 ka. Thus, ~190 ka is viewed as being a maximum age for the exposure of Liang Bua and a maximum age of occupation for the site. This chronological evidence provides a context for archaeolgical evidence found within these deep cave sediments.