Engel, M.

Physical experiments and numerical models support investigations on the transport of boulders by tsunamis, and based on this, on hydrodynamic characteristics of the tsunami itself. We conduct physical experiments in a flume applying idealized (cuboid) as well as a naturally shaped boulder, the latter representing a downscaled model from a field site on the island of Bonaire (Lesser Antilles, Caribbean Sea). Besides the boulder shape, we study the influence of shoreline morphology and pre-transport setting on boulder transport by a tsunami (Oetjen et al., 2020).

The physical experiments show that the interaction between bore and boulder differ significantly for the idealized and natural boulders. In our experiments, the natural boulder model consists of a lower drag coefficient, leading to lower flow disturbances and transport distances, and an increased entrainment threshold compared to the cuboid boulder. Subsequently, the natural boulder is thus transported approximately 30 % shorter than the cuboid one of same volume and weight. Since idealized shapes like cuboids are non-existent in nature, the results indicate that existing equations predicting entrainment thresholds or transport distances, overestimate the actual values. However, it is not clear to which amount the influence of the boulder shape is superimposed by other boulder and wave properties (e.g., ratio between wave velocity and boulder volume or weight) or local conditions (e.g., initial boulder submergence).

Furthermore, especially for experimental setups leading to high transport distances, significant fluctuations, of the transport distances are observed (up to 650 % in a single setup). This shows the sensitivity and complexity of coastal boulder transport and clarifies, that evaluating such processes in nature need to be conducted as accurate as possible while attributing to the large uncertainties associated with the transport process which might not be solvable for particular events (e.g., due to remobilization processes or unknown transport mode).

For most transport properties our findings are in line with previous studies. However, in contrast to some of them, we only observe sliding transport and higher variations in the transport distances. A large percentage of the deviations between our results and other studies, may also be related to divergent experimental setups, especially in terms of wave - boulder property ratios (e.g., increased ratios between boulder density and wave height/velocity).

Subsequently, comparisons between the findings of different studies are not straightforward. For simplifying this, we suppose a more coordinated research approach based on a standardized experimental setup. Such a setup would allow research to focus on single parameters and an easier comparison of results from other research groups, flumes and experimental campaigns.

Oetjen, J., Engel, M., Pudasaini, S.P., and Schuettrumpf, H.: Significance of boulder shape, shoreline configuration and pre‐transport setting for the transport of boulders by tsunamis. Earth Surf. Proc. Landforms, 45, 2118–2133, 2020, https://doi.org/10.1002/esp.4870.

The Caribbean region is highly vulnerable to coastal hazards since a relatively high percentage of the population lives right at the coast. Tourism, a major economic factor on many islands, is concentrated in coastal areas as well. The “traditional” hazards along the coasts of the Caribbean, as perceived some decades ago, are earthquakes, volcanism and storm surges during hurricanes. This focus is justified in the light of the devastating earthquake of Haiti in 2010, which represents a regional maximum in magnitude with a death toll of more than 230,000. The eruption of Mount Pelée on Martinique, which destroyed St. Pierre, the former principal town of the island, and killed around 28,000 inhabitants, and the Great Hurri-cane of 1780 with a similar number of fatalities along the Antilles island arc were outstanding disasters in terms of magnitude. Moreover, the frequency during the 500 years of historical documentation is high as well.
However, history tells that the Caribbean is also prone to the risk of tsunamis. One hundred twenty-seven possible tsunamis were documented during the last 500 years, of which 53 were finally considered to be real tsunamis.

Tsunami hazard on Bonaire
Since we found geological traces of potential tsunamis at almost all sites investigated along the coast of Bonaire, we consider the entire coastline to be prone to tsunami hazard. Along the windward coast, the largest boulders point to the occurrence of tsunami wave heights in the order of 8-10 m. This height estimate is similar to tsunami wave heights observed at the Venezuelan coast in historical times, e.g. at Paria in 1530 or at Puerto Tuy in 1900 [5]. The minimum inundation of such a tsunami on the carbonate platform along the windward side would be 300 m, possibly even more than 500 m. However, findings of candidate tsunami deposits in the sediment cores of the bays, bokas and saliñas, such as Boka Bartol, Saliña Tam or Lagun far inland indicate that especially these low-lying areas provide pathways for inundation and destruction by a tsunami. Thus, a low topography, as found on the entire southern part of Bonaire, is more prone to tsunami inundation.

Dating the transport/deposition time of supratidal coarse-clast deposits is difficult, limiting their value for inferring frequency-magnitude patterns of high-energy wave events. On Bonaire (Leeward Antilles, Caribbean), these deposits form prominent landforms, and transport by one or several Holocene tsunamis is assumed at least for the largest clasts. Although a large dataset of 14C and electron spin resonance (ESR) ages is available for major coral rubble ridges and ramparts, it is still debated whether these data reflect the timing of major events, and how these datasets are biased by the reworking of coral fragments. As an attempt to overcome the current challenges for dating the dislocation of singular boulders, three distinct dating methods are implemented and compared: (i) 14C dating of boring bivalves attached to the boulders; (ii) 230Th/U dating of post-depositional, secondary calcite flowstone and subaerial microbialites at the underside of the boulders; and (iii) surface exposure dating of overturned boulders via 36Cl concentration measurements in corals. Approaches (ii) and (iii) have never been applied to coastal boulder deposits so far. The three 14C age estimates are older than 40 ka, i.e. most probably beyond the applicability of the method, which is attributed to post-depositional diagenetic processes, shedding doubt on the usefulness of this method in the local context. The remarkably convergent 230Th/U ages, all pointing to the Late Holocene period (1.0–1.6 ka), are minimum ages for the transport event(s). The microbialite sample yields an age of 1.23 ± 0.23 ka and both flowstone samples are in stratigraphic order: the older (onset of carbonate precipitation) and younger flowstone layers yield ages of 1.59 ± 0.03 and 1.23 ± 0.03 ka, respectively. Four coral samples collected from the topside of overturned boulders yielded similar 36Cl concentration measurements. However, the computed ages are affected by large uncertainties, mostly due to the high natural chlorine concentration. After correction for the inherited component and chemical denudation since platform emergence (inducing additional uncertainty), the calculated 36Cl ages cluster between 2.5 ± 1.3 and 3.0 ± 1.4 ka for three of four boulders whilst the fourth one yields an age of 6.1 ± 1.8 ka, probably related to a higher inheritance. These 230Th/U and 36Cl age estimates are coherent with a suggested tsunami age of < 3.3 ka obtained from the investigation of allochthonous shell horizons in sediment cores of northwestern Bonaire. Whilst 230Th/U dating of post-depositional calcite flowstone appears to be the most robust and/or accurate approach, these results illustrate the potential and current limitations of the applied methods for dating the dislocation of supralittoral boulders in carbonate-reef settings.

Abstract:

We present sediment cores from seven coastal geoarchives on Bonaire, southern Caribbean, containing layers of high- energy sedimentation. Tsunami deposition is inferred for some layers based on the presence of allochthonous reefal shells including articulated specimens and a high percentage of angular fragments, planktonic foraminiferal taxa and those from the deeper shelf (below storm wave base), basal unconformities and hiatuses of >1000 a, rip-up clasts, thin depositional sequences comprising basal traction carpets overlain by normally graded sand, a proximal sediment source (littoral) in the lower part of the deposit and a broad mixture (littoral, shelfal, terrestrial) in the upper part, and the lack of deposition during recent hurricane flooding. Several tsunami layers were precisely dated to 3300-3100 cal BP, whereas the record of further candidate tsunamis is more disjunct. Additional tsunami evidence is provided by the largest coastal boulders (up to 150 t; a-axis up to 10 m). 

Abstract:

A sediment record of three alluvial sites along the east- and northeast-oriented shore of Bonaire (Netherlands Antilles) provides evidence for the recurrence of several extraordinary wave impacts during the Holocene. The interpretation of onshore high-energy wave deposits is controversially discussed in recent sedimentary research. However, it represents a powerful tool to evaluate the hazard of tsunami and severe storms where historical documentation is short and/or fragmentary. A facies model was established based on sedimentary and geochemical characteristics as well as the assemblage and state of preservation of shells and shell fragments. Radiocarbon data and the comparison of the facies model with both recent local hurricane deposits and global “tsunami signature types” point to the occurrence of three major wave events around 3300, 2000–1700 and shortly before 500 BP. Since (i) the stratigraphically correlated sand layers fulfill several sedimentary characteristics commonly associated with tsunamis and (ii) modern strong hurricanes left only little or even no sediment in the study areas, they were interpreted as tsunamigenic. However, surges largely exceeding the energy of those accompanying modern hurricanes in the southern Caribbean cannot entirely be ruled out. The results are partially consistent with existing chronologies for Holocene extreme wave events deduced from supralittoral coarse-clast deposits on Aruba, Bonaire and Curaçao as well as overwash sediments from Cayo Sal, Venezuela.