In our field survey, we recognized extensive tsunami inundation by the presence of debris and watermarks around the mouth of the Ukai River (Fig. 1c). However, the tsunami deposits were very limited in distribution, occurring as patches on the scale of hundreds of meters scale rather than being widespread along the coast.
Observed inundation height, debris, and liquefaction features around the study site
The height and direction of watermarks and debris (human-produced objects and garbage and drifting plants) allowed us to estimate the flow depth and direction of the tsunami waves. Watermarks around transect A–B (Fig. 1c) were 45–47 cm above ground level (Figs. 2a–c & Table 1). Plant debris trapped on trees at locations M and Q in Fig. 1c showed inundation heights of 140 cm and 147 cm above ground level, respectively (Fig. 2d & Table 1). Debris was observed up to about 160 m inland from the shoreline (Figs. 1d, 2f & Table 1). Plant material trapped by trees and poles near the coast (Fig. 2e & Table 1) was dragged in the seaward direction. In situ leaning plants in the paddy fields in the study area also indicated the seaward direction of the tsunami return flow. We observed obvious liquefaction features at many locations in the paddy fields (Fig. 2g & Table 1) near our transect. The diameters of vented liquefied sand areas ranged from 30 to 135 cm.
Thickness of the 2024 tsunami deposits
The thickness of the tsunami deposits along transect A′–A″ (0–48 m inland from the seawall) ranged from 0.6 to 11.2 cm (Figs. 2h & 3). Maximum and minimum thicknesses were recorded at 8 and 33.5 m inland from the seawall, respectively. Along the transect, horizonal change in thickness of the tsunami deposits fluctuated and showed a multimodal distribution (Fig. 3). Discontinuous faint sand sheets were observed on the paved road inland of the sampling sites. Tsunami deposits were also observed in a coastal park about 150 m south of the study site (Fig. 2i & Table 1), but these deposits occurred only sparsely and we did not measure changes in their thickness.
Stratigraphy
Location SZ1
Location SZ1, the most seaward sampling point of this study, was on a footpath. At this location, the tsunami deposit was 9.5 cm thick and consisted of five units (Units 1–5 from bottom to top, Fig. 4). During the field survey, we defined only two units, the uppermost part of the succession (Unit 5) and the lower part of the succession (all other units), on the basis of the sediment color (dark gray in Units 1–4 and brown in Unit 5). Five units were later distinguished based on their sedimentary structures in computed tomography (CT) and soft X-ray images. Unit 1 (7–9.5 cm from the top of the deposit), the lowest unit, was characterized by the presence of faint parallel laminations. Unit 2 (6–7 cm from the surface) was very thin and defined by a relatively high mud content and the presence of parallel laminations and hummocky cross-stratification (HCS) mimics32. Unit 3 (2–6 cm from the surface) was the middle part of the tsunami deposit and contained prominent climbing ripples with their leesides to seaward. The tops of the climbing ripples in this unit appeared to have been dragged and truncated seaward by the overlying Unit 5. Grain-size analysis showed inverse grading within Unit 3 (Fig. 4 & Supplementary Table S1). Unit 4 was recognized as coarse, poorly sorted sand within depressions at the top of Unit 3. Unit 4 contained many shell fragments and sponge spicules. The top of the tsunami deposit (Unit 5; 0–2 cm from the surface) was characterized by parallel laminations. This unit showed upward grading and a lower sharp contact.
Location SZ2
At location SZ2, which was in a small cultivated field, tsunami deposits 5.9 cm thick overlay the organic soil of the original ground surface (Fig. 5a). This unstructured tsunami deposit was divided on the basis of sediment color into an upper brown part (0–3 cm from the top of the deposit) and a lower dark gray part (3–5.9 cm from the top of the deposit). The upper part was probably correlated with Unit 5 at location SZ1, but the lower part of the sediments did not obviously correspond to the units at SZ1. The sediment samples from SZ2 were composed of fine sand (2.33–2.66 phi, average 2.50 phi), and multiple instances of normal grading were detected in the tsunami deposit (Fig. 5a & Supplementary Table S1). The basal contact of the tsunami deposit was very clear both by visual observation and in X-ray images, but the underlying soil did not appear to have been eroded.
Location SZ3
The appearance of the tsunami deposit at location SZ3 was similar to that at SZ2. At SZ3, the tsunami deposit was 5.6 cm thick (Fig. 5b) and was divided into an upper brown part (0–2 cm from the top of the deposit) and a lower gray part (2–5.6 cm) by visual observation in the field and laboratory. The tsunami deposit was composed of fine sand (2.39–2.78 phi, average 2.63 phi, Fig. 5b & Supplementary Table S1). The grain-size analysis results also revealed multiple grading in grain size within the tsunami deposit, as at location SZ2. These observations suggest that the upper brown part at this location is correlated with Unit 5 at locations SZ1 and SZ2. Parallel laminations were recognized in the upper brown part of the tsunami deposit by X-ray images.
A striking feature of the tsunami deposit at SZ3 was the mud content and grain size of the sand fractions at the bottom of the deposit. The lowermost sample exhibited the highest mud content (14.78%), lowest mean grain size, and highest sorting (0.84 phi) of the sand fractions. The basal contact with the underlying soil was distinct but not erosional.
Location SZ4
The 4.9-cm-thick tsunami deposit at SZ4 overlay the organic soil of the original ground surface (Fig. 5c). The tsunami deposit was divided during the field survey into an upper brown part (0–2 cm from the top of the deposit) and a lower dark gray part (2–4.9 cm) based on sediment color. The upper part probably correlated with Unit 5 at location SZ1. The grain size of the tsunami deposit ranged from 2.54 to 2.77 phi (average 2.67 phi, Fig. 5c & Supplementary Table 1). Grain-size analyses also showed multiple normal grading in the tsunami deposit. Parallel- to cross-lamination was visible in the upper part of the tsunami deposit in both soft X-ray and CT images. The mud content decreased from bottom to top of the tsunami deposit, but was exceptionally high near the top of the deposit. The contact of the tsunami deposit with the underlying soil was sharp.
Location SZ5
Location SZ5, the most landward sampling location, a tsunami deposit 3.7 cm thick overlay the organic soil (Fig. 5d). The deposit consisted only of dark gray sand. Parallel laminations were observed on CT and X-ray images around the top of the deposit (0 to ca. 2 cm below the top of the deposit). The mean grain size of the top of the deposit was greater than 3.00 phi, the range of very fine sand, and the mean grain size of the other samples ranged in fine sand (2.61–2.75 phi). Grain-size analyses showed a single normal grading in the sand fraction of the tsunami deposit (Fig. 5d & Supplementary Table 1). The mud content was highest at the top. The contact of the tsunami deposit with the underlying soil was distinct but not erosional.
Diatom assemblages within the 2024 tsunami deposit and reference samples
Diatom assemblages in 11 samples (U1-1 to U5-2 from bottom to top in Fig. 6d) from the tsunami deposits at location SZ1 were dominated mostly by brackish–marine and marine taxa, except for sample U4-1 (Figs. 6a & 6d). A brackish species, Catenula adhaerens, was the dominant species in the nine samples from Units 1–3, and 5 (samples U1-1 to U3-4 and sample U5-1). C. adhaerens was still abundant in the sample from the top of the deposit (U5-2), but the first dominant taxon in this sample was marine Delphineis spp. Sample U4-1 was characterized by a mixed assemblage of freshwater terrestrial (Luticola spp., Hantzschia amphioxys, and Stauroneis obtusa), brackish–marine, and marine taxa (Fig. 6a).
The diatom taxa found in the tsunami deposits were also common in samples from the underlying field soil and from the modern beach near our study site (Figs. 6). A field soil sample was dominated by freshwater terrestrial taxa, such as Luticola spp., Hantzschia amphioxys, and Stauroneis obtusa. The modern beach sand was dominated by brackish–marine and marine taxa (Catenula adhaerens and Delphineis spp.) and included Fallacia florinae, Achnanthaceae sp., Amphora sp. 1, and Cocconeis sp. 1–4 as accompanying taxa (Figs. 6a & 6b).
Samples of vented sand and paddy soils obtained ~ 120 m inland from the sampling sites were dominated by freshwater species (Figs. 6b & 6c). The relative abundance of Sellaphora sp. was more 50% in one of the samples from the vented sediments (V1). The other sample from the vented sand (V2) and paddy soils were characterized by Gomphonema spp., Placoneis undulata, Caloneis bacillum, and Sellaphora pupula complex.
Discussion and Conclusion
The 2024 tsunami inundation extended about 300–400 m inland from the shoreline in the area around the river mouth of the Ukai River, but the extent of the tsunami deposit was not widespread and limited in specific areas near the shoreline. Recent tsunami deposits associated with giant tsunamis (e.g., the 2004 Sumatra and the 2011 Tohoku events), are generally widespread and thin landward on scales of hundreds of meters or kilometers33–38; these distributional features can be explained by the decreased current velocity and sediment capacity of a tsunami wave as it moves inland39. Thus, the limited distribution of the tsunami deposit in our transect may be ascribed to small-scale changes in topography (e.g., mounts and depressions)35,40–43, topographic irregularities, and obstacles that caused quick changes in the velocity and direction of the current44,45, and irregular source of the sediment (e.g., vented sediments, as discussed below)46,47.
In X-ray CT and soft X-ray images of the tsunami deposit associated with the 2024 earthquake, the substantial structures recognized at location SZ1, the most seaward location, can be used to estimate the current direction and flow condition. The climbing ripples with their leeside to seaward in Unit 3 (Fig. 4) indicate transport and deposition of sediments under a seaward current. Although the parallel laminations and hummocky cross-stratification mimics observed in the basal Units 1 and 2 cannot be used to reconstruct current direction, the structures in Units 1–3 likely formed sequentially under a single seaward current because no clear erosional boundaries separate these three units. The inverse grading of the sand in Units 1–3 suggest that these three units formed successively as the current velocity increased. The occurrence of truncated and deformed climbing ripples at the boundary between Unit 3 and the overlying Unit 5 suggests that Unit 5 was deposited under a different seaward current from Units 1–3. The later seaward flow eroded the uppermost part of the climbing ripples and formed flat-head features at the top of the ripples. Immediately after deposition, the climbing ripples would have been ductile because of a high porewater content; thus the current could have dragged the underlying sediment seaward. The presence of upward normal grading in Unit 5 indicates decreasing flow velocity toward the final phase of deposition after the dragging and truncation of the ripples. These dragged and truncated structures have been reported in the 2004 tsunami deposit at Bang Sak, Thailand48. In addition, a model of accumulation of inverse and normal grading structures across an erosion surface during increasing and then decreasing flow velocity has been proposed for deposits formed by the 2004 and 2011 tsunamis49,50.
The presence of exotic sediment within the ripple troughs (Unit 4 at location SZ1) may indicate that there was another tsunami wave between the depositions of Units 3 and 5. The sediment samples from Unit 4 were poorly sorted and coarser than the sediments above and below and contained a mixed (terrestrial, brackish, and marine) diatom assemblage, shell fragments, and sponge spicules, whereas samples from Units 3 and 5 consisted of well-sorted fine sand and contained mostly brackish–marine diatom assemblages with low species diversity. This contrast may indicate that Unit 4 was generated by a different source (wave) from Units 3 and 5. Alternatively, Unit 4 may have originated as part of the initial supply of Unit 5; most of the sediments deposited during the early stage may have been subsequently eroded as the velocity of the seaward current increased. In either case, our observations do not allow us to determine exactly which current caused the initial deformation of the climbing ripples, or when this deformation took place.
The tsunami deposits at locations SZ2–SZ4 may also have been formed by multiple waves, as indicated by the presence of two normal grading structures (Figs. 5a–5c). The boundary between the upper and lower graded structures did not show clear erosion, but the basal part of the upper normal grading structure corresponded to the boundary between Unit 5 (brown part) and the dark gray part at location SZ1. Therefore, two waves with different mineral compositions probably generated the lower dark gray part and the upper brown part. Normal grading within the lower dark gray part, recognized in the tsunami deposits at SZ2–SZ4 and in the deposit at SZ5 (Figs. 5a–5c), may indicate transport and deposition of sediments by the same wave.
Diatom assemblages provide information on the sources of sediments within tsunami deposits. To understand the relationships among the assemblages, detrended correspondence analysis (DCA) was carried out using diatom assemblages from tsunami deposits. The DCA also included samples from underlying field soil, modern beach sand, vented sediments in paddy fields, and paddy soil, as reference for sediment source (Figs. 6e & 6f). The DCA ordination plots represent assemblage samples as points in multi-dimensional space. In the ordination, similar assemblages are located close together and dissimilar assemblages further apart51. Except for one sample from Unit 4, all of the diatom assemblages within the tsunami deposits were very similar. The compositions of the diatom assemblages from Units 1–3 and 5 were more similar to those of modern beach sand than to those of underlying field soil, whereas the sample from Unit 4 contained an assemblage similar to that of the underlying field soil (Fig. 6e & 6f). This difference probably reflects a difference in source between Unit 4 and the other tsunami samples. From such results, it might be natural to conclude that Unit 4 and the other units were formed by different waves and from different sources. However, in the tsunami deposits on Phra Thong Island, Thailand, associated with the 2004 Sumatra tsunami, diatom assemblages with different compositions in the upper, middle, and lower portions of a single tsunami deposit have been interpreted to have resulted from changes in current velocity52. Thus, on the basis of the micropaleontological evidence, we cannot determine with certainty whether Units 4 and 5 were formed by different waves with different sources, or by different stages of the same wave.
There are several possible sediment sources for the tsunami deposits: the beach, the pre-tsunami ground surface, and the riverbed of the Ukai Rive. The diatom assemblages suggest that the sediment sources of the tsunami deposits were brackish–marine (Units 1–5) and partially terrestrial (Unit 4) environments. The beach and seafloor are brackish–marine environments, and terrestrial environments include the pre-tsunami ground surface and the riverbed over which the tsunami wave passed.
In this study, liquefaction-induced sediments vented in paddy fields from the sand underneath are unlikely to be a source of the tsunami deposits, although liquefied vented sediments have sometimes been reported as the source of tsunami deposits (e.g., the 2011 tsunami in Sendai, Japan47, and an event about 1100 years ago in Puget Sound, USA46). In contrast to the samples from the tsunami deposit, which were dominated by brackish–marine diatoms, two samples from the vented sand were dominated by freshwater taxa that originated from the source of the liquefied sand or were ripped from the paddy soils during venting. The DCA ordination plots represented quite different compositions of diatom taxa among samples of tsunami deposits and vented sediments (Figs. 6e & 6f). In addition, grain-size analysis showed that the mean grain sizes of the vented sediments are remarkably coarser than any part of our tsunami deposits (Figs. 4, 5 & Table 2). We should also note that our data only show vented sediments from paddy fields, about 120 m west of the sampled locations. If vented sediments appeared near the coast along the transect during the time between the earthquake and the associated tsunami arrival, they could be the origin of the tsunami deposit, although eyewitness accounts of liquefaction prior to the tsunami arrival have not yet been found at this stage.
Table 2
The results of grain-size analysis on beach sand, vented sediments, and paddy soil.
Sapmle | Median [phi] | Mean [phi] | Sorting | Mud content [%] |
Beach sand | 2.39 | 2.38 | 0.45 | 5.2 |
Vented sediments (V1) | 2.03 | 2.03 | 0.48 | 3.1 |
Vented sediments (V2) | 1.87 | 1.84 | 0.58 | 4.99 |
Paddy field soil | 1.65 | 1.56 | 1.22 | 25.8 |
This study collected samples and documented the tsunami deposits near the mouth area of the Ukai River, Suzu City, within a few weeks of the 1 January 2024 Noto Peninsula earthquake. Our detailed observations of the tsunami deposits using soft X-ray and X-ray CT images and grain-size analysis revealed characteristic sedimentary structures, which indicate complex tsunami flow conditions and at least two seaward currents. Further, dragging and truncation of sediments were important parts of the formation of the tsunami deposit. The paleontological data allowed possible sources of the tsunami deposit to be estimated. This modern analogue provides criteria that can be used to identify tsunami deposits and to reconstruct the inundation areas, sedimentary processes, and recurrence intervals of past tsunamis.
Methods
Field survey
Fieldwork was carried out in January, February, and June 2024. The thickness of the tsunami deposit was measured at 0.5-m intervals at 96 locations up to 48 m from the seawall. For detailed observation of sedimentary structures and paleontological analysis, sediment samples were collected at five locations (SZ1–SZ5) by pushing a flat acrylic box into the wall of a small pit. Location SZ1 was on a path by the seawall; locations SZ2, 3, 4, and 5 were in a small field landward of the path. The locations for sample collection and observations and the heights of debris and watermarks were measured with a network Real Time Kinematic-Global Navigation Satellite System (RTK-GNSS) survey system from Leica Geosystems Inc. (Norcross, Georgia, USA).
Observation of sediments
We observed the sediment samples visually, and also obtained soft X-ray and X-ray computed tomography (CT) images. Soft X-ray and X-ray CT images were taken by using a SOFRON SRO-i503-2 (SOFTEX Co., Ltd., Tokyo, Japan) and with a Hitachi Supria Grande PREMIUM (Hitachi, Ltd., Tokyo, Japan), respectively, at the Geological Survey of Japan.
Grain-size analysis
Grain-size analysis was applied to the tsunami deposits at all five locations SZ1–SZ5, underlying field soil at the study site, beach sand, the vented sediments (location Y in Fig. 1c), and a paddy soil beneath the vented sediments (location X in Fig. 1c). Bulk samples were sieved using mesh cloth with 63 µm (4 phi) openings to measure the mud content. The grain-size distribution of sand particles was measured using an image analyzer (Camsizer, Retsch Technology GmbH, Haan, Germany). The mean, median, sorting, skewness, and kurtosis of the grain-size distributions were calculated using the logarithmic graphical method 53. The detailed results are shown in Supplementary Table S1.
Diatom analysis
Diatom assemblages within the tsunami deposits at SZ1 were identified and counted. We also identified the assemblages in underlying field soil at the study site, beach sand, vented sediments (location Y in Fig. 1c), and a paddy soil beneath the vented sediments (location X in Fig. 1c) as reference samples. Slides were prepared with a bleaching method54,55. For each slide, at least 200 diatom valves were identified under a light microscope at ×600 magnification. Diatom identification and ecological information followed standard and local studies56–65. Count data are provided in Supplementary Table S2. Detrended correspondence analysis (DCA) was performed with “vegan”66, a package written for the free open-source language R67. The script is provided in Supplementary Text S3.