Introduction

At 16:10 JST (JST = UTC + 9) on January 1, 2024, a large Mj7.6 (Mw7.5) earthquake occurred in the northern part of the Noto Peninsula, Japan (hereafter, referred to as the 2024 earthquake)1,2,3,4,5. It was preceded by an intense earthquake swarm that persisted for over three years. According to the Japan Meteorological Agency, at least ≥ 23,700 ≥ Mj1.0 earthquakes occurred between December 1, 2020 and January 1, 20246, including an Mj6.5 (Mw6.2) earthquake that occurred at 14:42 JST on May 5, 20237. Additionally, earlier in the record, an Mj6.9 (Mw6.7) earthquake occurred on March 25, 2007 in the northwestern part of the Noto Peninsula8,9,10. Although intense seismic activity has long jeopardized the Noto Peninsula region, the 2024 earthquake represents an extraordinary event in terms of related hazards and the aftermath. The 2024 earthquake caused widespread damage on the Noto Peninsula and triggered tsunamis that inundated eastern and western shorelines11,12,13. In the northeast of the Noto Peninsula, which was close to the epicenter, many landslides were generated that damaged buildings and roads12,14,15.

Earthquake source models have been used to locate the faults along the northern coast of the Noto Peninsula responsible for the 2024 earthquake3,4,5,12,13. The rupture of those faults probably triggered ground upheaval of the northern part of the Noto Peninsula from east to west, resulting in the uplift of coastal areas. Synthetic aperture radar (SAR) image analysis was used to estimate a maximum uplift of approximately 4 m around the northwestern part of the Noto Peninsula12,16. This 2024-earthquake-induced uplift instantaneously changed the regional coastal landforms and environments. The uplift resulted in shoreline advance along the northern Noto Peninsula, with corresponding expansion of the total area of the coastal plains12,13,14,15,16. Tectonic perturbations, coupled with natural succession and artificial activities (e.g., dredging and construction), can substantially affect coastal ecosystems and landforms. For example, following an intense storm that occurred over the Noto Peninsula on September 20–21, 2024, many landslides and floods were triggered that transported large amounts of sediment to coastal areas, which substantially reshaped the coastal landforms17. Accordingly, following the occurrence of earthquakes, accurate measurements of uplift-induced shoreline advances are necessary to investigate both their immediate and their long-term effects.

To quantitatively measure uplift-induced shoreline advance, we compared aerial photographs taken before and after the 2024 earthquake and focused on two bays where the largest uplift was estimated to have occurred. An orthophoto taken before the 2024 earthquake and one produced using structure from motion multiview stereo (SfM–MVS) photogrammetry from aerial photographs obtained by the Geospatial Information Authority of Japan (GSI)12 after the 2024 earthquake were analyzed. The goal of this study was to investigate the extent and trend of the shoreline advance and coastal area expansion of the two studied bays based on high-resolution (0.5-m intervals) measurements of shoreline differences before and after the 2024 earthquake.

Study area

Site description

We focused on Minazuki and Yoshiura bays in the northwest of the Noto Peninsula (Fig. 1a,b), which are separated by the Kozaki cape (Fig. 1c). On the basis of SAR image analysis, the maximum uplift around these bays, caused by the 2024 earthquake, was estimated at 4 m, and the maximum shoreline advance was approximately 200 m in Minazuki Bay12,16. Minazuki and Yoshiura bays are primarily underlain by Miocene sedimentary rocks (Fig. 1c)18, with some Holocene sedimentary rocks adjacent to the Minazuki Bay coast (Fig. 1c)18. Minazuki Bay has two fishing ports and Yoshiura Bay has one fishing port. Prior to the 2024 earthquake, the Minazuki and Igisu rivers drained into Minazuki Bay (Fig. 1c). The drainage areas of the Minazuki and Igisu rivers are 6.9 and 3.9 km2, respectively.

Fig. 1
figure 1

Study site: (a) overview, (b) Noto Peninsula, and (c) Minazuki and Yoshiura bays. Yellow circles in panels (a) and (b) indicate the locations of Minazuki and Yoshiura bays. Black star in panel (b) indicates the epicenter of the 2024 earthquake. Blue circle in panel (b) indicates the closest tidal gauge. Red lines in panel (b) indicate the locations of active faults, provided by the Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology, Japan24. Black polygons in panels (b) and (c) indicate earthquake-induced landslides identified by the Geospatial Information Authority of Japan12. Contours in panel (c) are at 10-m intervals. Inset in panel (c) depicts the geological setting18. All maps were generated using QGIS version 3.28 (http://qgis.org/).

Full size image

2024 Noto Peninsula earthquake

At 16:10 JST on January 1, 2024, an Mj7.6 (Mw7.5) earthquake occurred in the northeast of the Noto Peninsula, Japan1,2,3,4,5. The epicenter was located at 37.50°N, 137.27°E (Fig. 1b)4. Earthquakes have frequently occurred in the northern part of the Noto Peninsula since 20206,7. Migration of crustal fluids has been suggested as the driving force of this earthquake swarm19,20,21,22, which might be related to notable increase in excess pore pressure associated with intense snowfall23. On May 5, 2023, an Mj6.5 (Mw6.2) earthquake caused substantial damage, especially in the north of the Noto Peninsula7. It has been proposed that the earthquakes that occurred after 2020 facilitated the rupture of faults lying along the northern edge of the Noto Peninsula (Fig. 1b)24, which consequently triggered the earthquake on May 5, 20237. In a similar manner, the 2024 earthquake was also probably affected by these previous earthquakes.

The 2024 earthquake triggered tsunamis and landslides. The tsunamis caused severe damage to northern parts of the Noto Peninsula near the earthquake epicenter. The heights of these tsunamis were estimated to have been in the range of 1.0–4.4 m on the west of the Noto Peninsula13. A GSI investigation using aerial photographs estimated that at least 2,300 landslides were triggered by the 2024 earthquake, and that the total landslide area across the Noto Peninsula, including the sediment-inundation zone, was approximately 8.8 km212. The maximum landslide area exceeded 0.4 km2 (i.e., within the order of 105 m2)12.

Methods

Data and processing

We used orthophotos to investigate the 2024 earthquake-induced shoreline advance. Orthophotos taken before and after the 2024 earthquake were obtained using different methods. The pre-2024 earthquake orthophoto, processed from aerial images captured on September 12, 2020, had high spatial resolution (up to 0.25 m); therefore, it clearly showed almost all surface structures. Accordingly, we assumed that it was well georeferenced and indicated true coordinates.

On the basis of this assumption, the orthophoto taken after the 2024 earthquake was georeferenced to spatially fit the pre-2024 earthquake orthophoto. First, six aerial images captured on January 11, 2024 by GSI12 were processed following a standard SfM–MVS workflow25,26,27 using photogrammetry software (Agisoft, Metashape Professional) to produce the post-2024 earthquake orthophoto with spatial resolution of approximately 0.8 m. We then selected artificial structures (e.g., fishing ports and schools) that were clearly visible in both orthophotos and extracted the coordinates of their corners from the pre-2024 earthquake orthophoto as ground control points (GCPs). Using the georeferencer processing tool in QGIS 3.16, we manually selected points in the post-2024 earthquake orthophoto corresponding to their respective GCPs and georeferenced them to fit the pre-2024 earthquake orthophoto. The average GCP georeferencing error was approximately 0.7 pixels (i.e., approximately 0.6 m). This indicated that comparisons between the orthophotos would accurately detect two-dimensional uplift-induced shoreline advance because most shoreline deviation was on a scale greater than tens of meters.

Tectonic perturbation caused by the 2024 earthquake induced both vertical uplift and horizontal ground displacement. Generally, SfM–MVS photogrammetry can accurately represent surface morphology, allowing investigation of three-dimensional displacement. However, the surface of water bodies (i.e., sea and rivers) typically changes in response to waves and turbulence, resulting in variations in brightness and color among captured images depending on the direction of shooting. These differences among the captured images impede accurate representation of water surfaces using SfM–MVS photogrammetry, leading to distortion of surface morphology near water bodies. Indeed, in flume tests investigating mixtures of granular material and water, designed to imitate debris flows, the morphological data processed using SfM–MVS photogrammetry failed to represent accurately either the fluid or nearby areas28. Similarly, in this study, distortion of the sea surface led to unavoidable errors in representing foreshore morphology. Owing to limitations of the dataset, this study did not delve into the vertical displacement caused by ground upheaval; instead, it considered only the relative horizontal changes in the shoreline before and after the 2024 earthquake. Georeferencing was performed only to align the pre- and post-2024 earthquake orthophotos horizontally, assuming uniform horizontal movement (i.e., rigid block-type motion) within the study area.

Aerial images of damage were urgently required following the 2024 earthquake12, and therefore some of the aerial images acquired were suboptimal for SfM–MVS photogrammetry (e.g., cloud cover or low overlap ratio). This was the technical reason why the study area was limited to Minazuki and Yoshiura bays.

Shoreline analysis

To investigate shoreline advance, the shoreline of each bay before and after the 2024 earthquake was drawn using the orthophotos. Shorelines before and after the 2024 earthquake were manually traced as lines along beaches and around fishing ports, caissons, and rubble breakwaters. Considering the high spatial resolution and good quality of the pre-2024 earthquake orthophoto, the traced pre-2024 earthquake shorelines were expected to be highly accurate. In the post-2024 earthquake orthophoto, the expanded coastal plains were bright, whereas the sea surface was dark. This contrast, reflecting the uplift of the underlying rocks, produced a clear boundary between the coastal plains and the sea, allowing accurate tracing of the post-2024 earthquake shoreline. Therefore, we assumed that the traced pre- and post-2024 earthquake shorelines were of similar accuracy and were thus comparable.

It should be noted that tidal correction of the extracted shorelines is necessary to strictly analyze coastal changes owing to shoreline advance29. In Japan, the tidal range varies from 1 to 2 m on the Pacific coast, but it has only a narrow range of approximately ± 0.3 m in the Sea of Japan (i.e., in the examined bays)29. The magnitude of tidal range in the Sea of Japan is considerably smaller than that of the reported surface deformation; therefore, the effects of the tide were considered negligible in this study. To investigate the tidal influence in detail, we confirmed the astronomical tide based on the closest tidal gauge (Fig. 1b). The post-2024 earthquake shoreline was traced from the orthophoto of approximately 11:00 JST on January 11, 2024, when the sea level was approximately 0 m. At the time of acquisition of the source aerial images used to create the pre-2024 earthquake orthophoto (i.e., September 12, 2020, 11:00–14:00 JST), the average sea level was relatively high at approximately 0.3 m. For the pre-2024 earthquake shoreline, some sections were traced along near vertical edges of artificial structures (e.g., fishing ports and caissons), meaning that the impact of such a high sea level on tracing the shoreline was likely only slight.

Using the traced shorelines, we investigated the extent of the shoreline advance. First, measurement points were placed along the traced pre- and post-2024 earthquake shorelines at 0.5-m intervals. Then, distances between the post-2024 earthquake measurement points and the closest measurement points along the shoreline before the 2024 earthquake were measured. These measured distances represented the extension of the shoreline, i.e., the extension distance was defined as the length of the shortest 0.5-m-interval lines connecting the shorelines before and after the 2024 earthquake. The expanded area of the coastal plain was measured as the area encompassed by the shorelines before and after the 2024 earthquake. We then compared the characteristics of the measured extension distances and the expanded areas of Minazuki and Yoshiura bays.

Results

The 2024 earthquake induced shoreline advance of up to approximately 200 m. In Minazuki Bay, shoreline extension occurred in almost all sections of the bay, but it was particularly substantial in the midsection (Fig. 2). The shoreline extension ranged between 0 and approximately 197 m (Fig. 2b), resulting in increase in the area of the coastal plain by 0.33 km2 (Fig. 2a). The shoreline advance also extended the total length of the shoreline of approximately 310 m. Consequently, the Minazuki and Igisu fishing ports dried up (Fig. 2).

Fig. 2
figure 2

Orthophotos of Minazuki Bay (a) before and (b) after the 2024 earthquake. Blue and red lines in panel (a) indicate the shoreline before and after the 2024 earthquake, respectively. Line colored from purple to yellow in panel (b) indicates the extension distance of the shoreline owing to uplift-induced shoreline advance. White arrows in panel (b) indicate the shooting direction of the photos shown in Fig. 4.

Full size image

In Yoshiura Bay, the extent of the shoreline advance was similar to that in Minazuki Bay. Shoreline extension occurred in most sections of the bay, with the exception of the rocky shore areas at both ends of the bay, expanding the area of the coastal plain by 0.13 km2 (Fig. 3a). The maximum extension distance of the shoreline was approximately 146 m, which occurred in the midsection of the bay (Fig. 3b). This maximum was less than that in Minazuki Bay (Figs. 2b and 3b). Nevertheless, the Yoshiura fishing port also dried up owing to this shoreline extension (Fig. 3), indicating that the changes in the Yoshiura Bay coastal landforms were also remarkable. The shoreline advance extended the total length of the shoreline by approximately 67 m.

Fig. 3
figure 3

Orthophotos of Yoshiura Bay (a) before and (b) after the 2024 earthquake. Blue and red lines in panel (a) indicate the shoreline before and after the 2024 earthquake, respectively. Line colored from purple to yellow in panel (b) represents the extension distance of the shoreline owing to uplift-induced shoreline advance.

Full size image

Photographs of Minazuki Bay support the changes inferred from interpretation of the orthophotos. A panoramic photo of Minazuki Bay taken before the 2024 earthquake shows two rubble breakwaters: one in the center of the bay and the other on the left-hand side (Fig. 4a). After the shoreline advance, the center breakwater was located on the new shoreline and consequently integrated into the coastal plain (Fig. 4b). It is evident from panoramic photos taken before and after the 2024 earthquake that the Igisu fishing port, surrounded by the left-hand breakwater (Fig. 4a,b), completely dried up owing to the shoreline advance and became part of the coastal plain (Fig. 4c,d). Additionally, the Minazuki fishing port on the opposite side of the bay also dried up following the 2024 earthquake (Fig. 4e). It has been estimated that the uplift in the Igisu and Minazuki fishing ports reached 4.1 and 3.3 m, respectively, based on a previous sea surface trace30.

Fig. 4
figure 4

Photographs of Minazuki Bay. Panels (a) and (b) show Igisu fishing port before (November 22, 2020) and after (January 17, 2024) the 2024 earthquake. Panels (c) and (d) show Igisu fishing port after the 2024 earthquake (January 4, 2024). Panel (e) shows Minazuki fishing port after the 2024 earthquake (January 18, 2024). Uplift heights shown in panels (d) and (e) are based on a previous sea surface trace30.

Full size image

The shoreline advance reshaped the morphology of the two river mouths inside Minazuki Bay. Before the 2024 earthquake, a sandy beach connected the Minazuki River and the shoreline (Fig. 5a). After the shoreline advance, the Minazuki River was forced to flow over the uplifted extended coastal plain (Fig. 5b). The uplifted surface of the new coastal plain consisted of not only beach sand but also newly emerged rocks, which were often captured as dark colors in the pre-2024 earthquake orthophoto (Fig. 5a,b). After the shoreline advance, the Minazuki River flowed across this uplifted rocky area, which is regarded as a shore platform (Fig. 5a,b), and a new sandy beach formed in the vicinity of the lower end (Fig. 5b). Consequently, the Minazuki River was diverted toward its right bank (Fig. 5b–d). This morphological change of the river mouth extended the total length of the Minazuki River by approximately 282 m (Fig. 5b), i.e., greater than the maximum distance of shoreline extension (Fig. 2b).

Fig. 5
figure 5

Changes around the Minazuki River mouth. Panel (a) is the pre-2024 earthquake orthophoto, panel (b) is the post-2024 earthquake orthophoto, and panels (c) and (d) are photos of the Minazuki River mouth after the 2024 earthquake (January 18, 2024).

Full size image

Similarly, there was a sandy beach connecting the Igisu River and the sea before the 2024 earthquake (Fig. 6a). Following the earthquake, the river mouth was shifted into the rocky area (Fig. 6b,c). Thus, the total length of the Igisu River was extended by approximately 121 m (Fig. 6b). In relation to this river extension, the extended channel eroded the sandy beach (Fig. 6c).

Fig. 6
figure 6

Changes around the Igisu River mouth. Panel (a) is the pre-2024 earthquake orthophoto, panel (b) is the post-2024 earthquake orthophoto, and panel (c) is a photo of the Igisu River mouth after the 2024 earthquake (January 18, 2024).

Full size image

Discussion

Uplift-induced shoreline advance extended the shorelines and consequently expanded the coastal plains of Minazuki and Yoshiura bays (Figs. 2, 3, 4, 5 and 6). The spatial extension of the shorelines tended to be greater in the midsection of both bays compared with that at the edges (Figs. 2b and 3b). This is potentially linked to the pre-earthquake coastal topography. In the midsection of both bays before the 2024 earthquake, sandy beaches connected the coastal plain to the sea (Figs. 2a and 3a), indicating that the shallow sea was the dominant shoreface topography. However, around the edges of the bays, rocky shores separated the coast from the sea (Figs. 2a and 3a). The sea depth around the rocky shore of the Kozaki cape, which separates Minazuki and Yoshiura bays, increases sharply seaward (Fig. 1c). In this context, the spatial variations in the shoreline extension distance likely correspond to variations in the pre-earthquake bathymetry, which reflect topographic differences between the sandy and rocky shores. In other words, the shoreline extension mainly reflects the topographic characteristics of the bays rather than localized uplift within or across the bays. This is supported by a statement from a fisherman who witnessed the entire area of Igisu fishing port being uplifted instantaneously by the 2024 earthquake30.

The three fishing ports (i.e., Minazuki, Igisu, and Yoshiura) are located at the edges of the bays, and therefore the shoreline extension distance around the ports is relatively small (Figs. 2 and 3). Nevertheless, owing to the shoreline advance, all three fishing ports dried up and are now inoperative (Figs. 2, 3 and 4). The shoreline advance hinders not only the revival of the local fisheries but also the shipping of earthquake relief supplies. Additionally, there were no reports of tsunami damage in either bay attributable to the 2024 earthquake11,30 and no such evidence was observed in the orthophotos (Figs. 2b and 3b). Theoretically, expansion of the coastal plain in combination with changes in shoreface topography and coastal landforms might have reduced the velocity and extent of any tsunami. However, to verify this inference, theoretical modeling of tsunami dynamics considering various uplift scenarios is required.

Measurements of changes in coastal landforms are likely affected by ocean tides. On the basis of the astronomical tide derived from the closest tidal gauge (approximately 0 m), the post-2024 earthquake shoreline can be approximated as the shoreline without the tidal effect. In contrast, the pre-2024 earthquake shoreline was measured under a relatively high sea level (approximately 0.3 m). The foreshore slope in the Sea of Japan ranges between 2.8° and 7.5°31. Given this range in foreshore slope, the difference in the post- and pre-2024 earthquake sea level (i.e., approximately 0.3 m) could cause errors of approximately 2.3–6.1 m in the horizontal measurement of the shoreline, which are minor compared with the overall scale of shoreline advance due to the 2024 earthquake (Figs. 2 and 3). The estimated horizontal errors (approximately 2.3–6.1 m) are comparable to the shoreline changes that followed the Mw9.0 2011 Tohoku earthquake (2.6–6.6 m/year), but are substantially less than the changes that occurred immediately after the same earthquake (i.e., shoreline advance of 20–40 m)32. The maximum advance of the shoreline of approximately 200 m, determined in this study, coincides with an estimate based on other data (i.e., SAR image)16, supporting the supposition that ocean tides likely had little impact on the examined bays. Hence, we conclude that the presented shorelines might have a maximum horizontal error of approximately 6 m, but that this can be ignored given the scale of the shoreline advance induced by the 2024 earthquake.

The 2024 earthquake reshaped the local coastal landforms such as sandy beaches, coastal plains, and shore platforms and consequently might have affected the beach and shoreface morphodynamics. The reshaped coastal plains might serve as a new sediment sink because of their relatively flat surface. The emergent coastal plains extended the course of both the Minazuki River and the Igisu River, and new sandy beaches developed at the mouths of the reshaped rivers. Sandy beaches are coastal landforms formed by cross-shore sediment transport between the lower and upper parts of the shoreface, balancing the local sediment budget of coastal areas33. However, in the basins of the Minazuki and Igisu rivers, earthquake-induced landslides produced a supply of readily mobilized sediment that could be transported to the shoreface in response to rainfall (Fig. 1c). Indeed, a rainstorm that occurred on September 20–21, 2024 transported a large amount of sediment to the river mouths, substantially modifying the shoreface morphology17. Therefore, coupled with landslide triggering, the 2024 earthquake altered the sediment budget controlling the shoreface morphodynamics through uplifted-induced shoreline advance, indicating that a long recovery time is anticipated for stable coastal landforms. In this context, the presented shorelines measured immediately after the 2024 earthquake could represent an indispensable benchmark for investigating future changes in coastal landforms, and contribute to elucidating the lasting effects of the 2024 earthquake on shoreface morphodynamics.

Uplift-formed shore platforms were also observed in coastal areas uplifted by the 2016 Kaikōura earthquake (Mw7.8) in New Zealand34,35. Denudation of shore platforms generally progresses via synergistic interplay between waves and weathering processes associated with various conditions (e.g., wet/dry cycles and wave energy)35,36,37,38. In particular, on the basis of field measurements and modeling of instances occurring on the Pacific Ocean side of Japan, it has been suggested that the lithological characteristics of shore platforms place important constraints on the extent of weathering and thus the rate of platform denudation39. In the examined bays, because the uplift-formed shore platforms comprise relatively hard sedimentary rocks (alternating layers of mudstone, sandstone, and conglomerate)40, the denudation rate is likely low compared with that of more friable lithologies (e.g., lapilli-tuffs)39. The marked lithological differences between Minazuki Bay (Holocene) and Yoshiura Bay (Miocene) (Fig. 1c) might be responsible for the variations in their denudation rates. Therefore, on the long-term scale of hundreds of years, the differences in the lithologies of the shore platforms would potentially lead to variation in the coastal landforms between the examined bays.

Conclusions

On January 1, 2024, an Mj7.6 (Mw7.5) earthquake occurred in the north of the Noto Peninsula, central Japan. The related displacement of ruptured faults triggered uplift of up to approximately 4 m along the coast, resulting in shoreline advance. To measure the extent and trend of the shoreline advance, we analyzed orthophotos taken before and after the earthquake, focusing on two bays in the northwest of the Noto Peninsula where the largest uplift occurred. The shoreline advance extended the shoreline by up to approximately 200 m in the midsection of both bays, whereas at the edges of the bays, shoreline advance was negligible. This indicates that the extent of the shoreline advance was constrained mainly by the previously existing shoreface topography and sea depth. Thus, the uplift likely occurred throughout both bays with less variation in magnitude from the north toward the south. The area of the coastal plains increased by 0.46 km2; thus, all three fishing ports in the bays dried up and became inoperative. Before the uplift, sandy beaches connected the rivers in each bay to the sea. However, the earthquake-induced uplift of undersea rocks formed new coastal plains, meaning that the previous sandy beaches no longer functioned as sediment sinks. This might have disturbed the balance of the sediment budget of the corresponding drainage basins, and a certain period might be required before stable coastal landforms are formed in the studied bays.