Post-seismic topographic shifts and delayed vegetation recovery in the epicentral area of the 2018 Mw 6.6 Hokkaido Eastern Iburi earthquake

Abstract

The 2018, M_w 6.6 Hokkaido Eastern Iburi earthquake in Japan triggered over 10,000 landsliding in an area spanning about 500 km², altering the local topography and leading to the accumulation of loose deposits on hillslopes and in valleys. However, a comprehensive post-seismic landslide inventory and an assessment of topographic changes are lacking, hindering a quantitative hazard assessment. Additionally, the extent of vegetation recovery in areas affected by coseismic landslides, a key indicator of post-seismic debris flow hazard, has not been evaluated. Here, we utilize high-resolution digital elevation models and multi-temporal satellite imagery to analyze topographic changes and vegetation dynamics in the earthquake’s epicentral area (seismic intensity >5.5). We observe that the event roughened the overall gentle topography of the region and made the slopes steeper. Owing to the absence of significant rainstorms and snowmelt post 2018, only a few debris remobilizations (60) and new landslides (80) have occurred in the affected region. Moreover, we noticed a slow vegetation recovery in the post-seismic phase, suggesting that the likelihood of debris flows and gully erosion remains elevated, highlighting the need for continued monitoring and assessment.

Keywords

Hokkaido Eastern Iburi earthquake LiDAR digital elevation models vegetation recovery post-seismic landslides topographic shift

Introduction

Major earthquakes can trigger widespread landsliding in steep orogens (Dou et al., 2020; Marc et al., 2015; Roback et al., 2018; Shao et al., 2019; Xu et al., 2014), resulting in long-lasting chains of geologic hazards (Bontemps et al., 2020; Fan et al., 2019). While the characteristics and patterns of coseismic landslides are relatively well studied, those of their remobilizations and the occurrence of post-seismic landslides, years to decades after the mainshock, are less constrained (Croissant et al., 2017; Fan et al., 2021; Keefer 1994; Marc et al., 2015). Rainfall events are the main trigger for remobilizations of coseismic landslide deposits (Bontemps et al., 2020; Domènech et al., 2019; Massey et al., 2014), as the affected topography, generally composed of destructured and unconsolidated materials, is often only marginally stable. The large availability of debris on steep slopes and in low-order channels determines an abundance of debris flow events for several years (Korup et al., 2007; Tang et al., 2009; Yunus et al., 2020). Additionally, the strong shaking creates fractures and weakens rock masses, particularly along ridges, resulting in enhanced weathering, erosion, and landsliding over extended periods (Fan et al., 2018c; Massey et al., 2014; Meunier et al., 2007; Parker et al., 2015).

Debris flow disasters and topographic changes have been reported systematically after major events such as the 1999 M_w 7.6 Chi-Chi and 2008 M_w 7.9 Wenchuan earthquakes. Typhoon Morakot in 2009 triggered over 10,000 landslides in southwest Taiwan (Steer et al., 2020), while Fan et al. (2018a) recorded more than 5000 remobilizations in 2011 near the epicenter of the Wenchuan earthquake. Dahlquist and West (2019) compiled an inventory of over a thousand debris flows associated with the 2015 Gorkha earthquake.

It is apparent that a better understanding of surface dynamics following strong earthquakes can help reduce or prevent disasters, in addition to providing clues on the role of earthquakes and mass wasting in shaping topography (Croissant et al., 2017; Dai et al., 2021; Fan et al., 2018a, 2021; Hovius et al., 2011; Marc et al., 2015; Yunus et al., 2023). In this regard, Shao et al. (2019) and other researchers attempted to produce quantitative hazard assessments of affected areas by developing accurate susceptibility models. Yet, studies that emphasize the role of strong earthquakes in denudation-erosion processes have received comparatively less attention because of a lack of pre-seismic topographic data. These data, where available, have been used to quantitatively estimate volumes, relief, slope, and roughness changes (Ren et al., 2014; Tang et al., 2019).

Multi-temporal digital elevation models (DEMs) and remote sensing images are powerful means to track topographic changes (Chen et al., 2006; Chigira and Yagi, 2006; Li et al., 2018; Ren et al., 2014; Shen et al., 2020). After the Wenchuan earthquake, Ren et al. (2014) and Li et al. (2018b) investigated topographic changes with the help of DEMs and found that the Wenchuan earthquake smoothed the steep relief and caused a co-seismic uplift of the Longmen Shan region. Furthermore, the topographic changes indicate that the greatest seismically induced denudation occurred in association with a thrust faulting mechanism and low-angle fault geometry. Using high-resolution aerial and satellite images, Fan et al. (2019) identified active landslide areas within the coseismic landslide boundaries. Vegetation recovery, as a proxy of slope stabilization, was also tracked over time (Yang et al., 2018a; Yunus et al., 2020). To this aim, the normalized difference vegetation index (NDVI), a dimensionless index that describes the density of greenness on a patch of land, extracted from multiphase remote sensing images was used, and indicators such as the vegetation damage area (VDA) and vegetation recovery rate (VRR) were obtained (Jiang et al., 2015; Lin et al., 2005, 2006).

The spatial pattern of vegetation phenology is an important factor that affects the stability of shallow landslides (Chen et al., 2020; Ren et al., 2014; Yang et al., 2018b). Korup et al. (2007) reported that extensive damage to the vegetation can affect the post-earthquake slope stability. Analyzing vegetation loss and recovery can be useful for quantifying potential landslide disasters after earthquakes (Saba et al., 2010). Numerous studies have confirmed that vegetation regrowth indicates a lack of landslide activity and possibly stabilization of landslide bodies (Lin et al., 2005; Yang et al., 2018a; Yunus et al., 2020). Based on vegetation regrowth trends, some researchers have predicted the time required for coseismic debris to stabilize (e.g., Chen et al., 2020; Xiang et al., 2023). After the Wenchuan earthquake, Yang et al. (2018b), Yunus et al. (2020), and others used the vegetation recovery rate to analyze the evolution of post-seismic landslides. In particular, Yang et al. (2018b) predicted that post-seismic landslide activity would return to pre-seismic conditions within two decades through an eight-year vegetation recovery analysis. Yunus et al. (2020) forecasted that landslide activity may return to the pre-earthquake level within 18 years based on the NDVI trend. In contrast, Gan et al. (2019), when analyzing the Mianyuan river basin, ∼80 km from the Wenchuan earthquake’s epicenter, predicted a faster recovery for vegetation (8–9 years). Evidently, the interplay between climate and topography in regulating vegetation equilibrium is complex, which warrants further in-depth investigation.

The 2018, M_w 6.6 Hokkaido Eastern Iburi earthquake in Japan triggered over 10,000 landslides (Dou et al., 2020). However, to date, the topographic changes and post-seismic landslide activity associated with this event have not been quantified. Here, we integrate two-time period LiDAR-derived high-resolution digital terrain models and satellite observations to study the detailed topographic changes caused by the earthquake, the evolution of vegetation recovery in landslide areas, and their implications for future hazard cascades. In most previous studies, coarser (25 m–90 m) pre-earthquake DEMs have been employed (Kober et al., 2015; Li et al., 2017; Ren et al., 2014), resulting in large uncertainties in the estimates of seismically induced denudation. Our results from two high-resolution LiDAR datasets and satellite analysis may reveal accurate denudation information, stability, and recovery time for the earthquake-affected area, thus proving useful for early warning of catastrophic debris flows and to understand the potential long-term impact of cascading hazards.

Study area

The M_w 6.6 earthquake that struck the Iburi sub-prefecture in southern Hokkaido on 6 September 2018 left 42 people killed, ∼750 injured, and 462 houses destroyed (Takahashi et al., 2020; Yamagishi and Yamazaki, 2018). The earthquake’s epicenter was located at 42.68°N, 141.93°E (Figure 1), and its focal depth was ∼37 km (JMA, 2018), indicating an inland earthquake caused by a blind fault adjoining the Ishikari-Teichi-Toen Fault zone. Asano and Iwata (2019) reported a maximum slip of 1.7 m at a depth of ∼26 km southwest of the hypocenter. The total seismic moment estimated through a fault-plane model was in the order of 10¹⁹ Nm (Kobayashi et al., 2019). Coseismic landslides occurred within an area of ∼512 km², where the relief varied between 0.06 and 505.57 m (average: ∼150 m). Low-elevation hills are the most common landform in the area, where slope angles of 15°–35° (average: 20°) are found in over 70% of the area (Dou et al., 2020).

The region mainly features deposits of Quaternary ash and pumice fallen from active volcanoes, underlain by sedimentary rocks from the Neogene (Nam and Wang, 2019; Yamagishi and Yamazaki, 2018). The surficial soil layers feature blanketed volcanic materials (4–5 m thick) from mount Tarumae (∼9000 years BP), Eniwa (∼20,000 years BP), and the Shikotsu (∼46,000 years BP) caldera volcanoes, and include pumice, volcanic ash, and clays (Kawamura et al., 2019; Yamagishi and Yamazaki 2018). Local faults strike from NNE-SSW to NNW-SSE (Kita 2019; Zang et al., 2019; Zhang et al., 2019). However, the 2018 event did not produce any observed surface rupture (Hua et al., 2019).

The area belongs to the temperate climatic zone. Precipitation in 2018 (Abira station, Hokkaido) was greater than that recorded in 2019 and 2020 (Figure S1, supplementary files). In fact, both the rainfall and snowfall during 2019 and 2020 were among the lowest in the last 30 years (Figure S1).

Data and methods

Remote sensing data

We obtained the pre-earthquake 10 m (vertical precision 2.5 m) DEM and 1 m (vertical precision 0.12 m) post-earthquake DEMs for the study area from the Geospatial Information Authority, Japan and Local Hokkaido Government, respectively. These two DEMs cover ∼512 km², that is, ∼95% of the earthquake affected area. Additionally, we obtained a 1 m pre-earthquake DEM from local government sources, which only covers 29 km². This small area DEM was used for calibration purpose. The vegetation phenology and time series post-earthquake landslide inventory in this study were mapped using Sentinel-2 images. Compared to MODIS (MODerate-resolution Imaging Spectroradiometer) and Landsat 8, Sentinel-2 offers superior resolution (10 m vs 250 m and 30 m, respectively) and frequent passes (∼5 days), which increase the availability of cloud-free products. For interpretation and validation of mapping accuracy, field photos, 3 m Planet images, and archived Google Earth images were used (Table 1).

Table 1.

Overview of the data used in this study. Note: all data cover the entire study area except for the 2012 LiDAR DEM that only covers 29 km².

Data	Acquisition time	Resolution(m)	Data source
DEM	Before 2018.9.5	10	Geospatial information authority (GSI)
LiDAR DEM	After 2018.9.5	1	Local Hokkaido government
LiDAR DEM	2012	1	Local Hokkaido government
Aerial photographs	2017.9.5	0.2	Geospatial information authority (GSI)
Aerial photographs	After 2018.9.5	0.2	Geospatial information authority (GSI)
Planet	2020.10.18	3	Planet labs
Sentinel-2	2017.6.17	10	European Space Agency (ESA) https://scihub.copernicus.eu/dhus/#/home
	2017.9.5
	2018.9.15
	2019.5.23
	2020.6.1
	2021.6.26
	2022.5.17

The 10 m resolution pre-earthquake DEM data were derived through topographic maps that cover the whole area of Japan published by GSI (https://fgd.gsi.go.jp/download/ref_dem.html), whereas the 1 m DEM was derived from LiDAR point clouds. Because the source, resolution, and processing steps are different between the two DEM products, we initially resampled the 1 m LiDAR DEM to 10 m resolution (i.e., the native resolution of GSI DEM) using cubic convolution to achieve a comparable performance. Further, to quantify the uncertainties and intrinsic errors between the two DEMs, we carried out a blank test in a non-landslide portion of the study area, adopting the mean error and R² as evaluation metrics (see Supplemental Figure S3). Notably, the R² was >0.97 in all cases, with a mean error of −3.20 m in relief. The results of the topographic change analysis were further validated with the 1 m pre- and post-earthquake LiDAR DEMs.

Coseismic and post-seismic landslide inventories

The coseismic landslide inventory was prepared by visually interpreting 0.2-m resolution aerial photographs, integrated with the 1 m LiDAR DEM to eliminate the shadow effect. A total of 10120 coseismic landslides (gray polygons in Figure 1) were delineated by manually interpreting the aerial photographs (Dou et al., 2020). According to the Geospatial Information Authority of Japan (GSI), the cloud cover and shadows produced by trees and ground relief hindered the identification of landslides in some locations. These issues were eliminated with the help of the LiDAR DEM. Additionally, in this inventory, the landslide scar and deposition areas were discriminated by visual interpretation of aerial photographs. The total mapped landslide area was 27.97 km². We cross-compared the aerial photographs from the pre-earthquake period (October, 2017) to ensure that all of the mapped landslides in the inventory could be associated with the M_w 6.6 Hokkaido Eastern Iburi earthquake. The new and remobilized landslide areas and vegetation recovery from 2019 to 2022 were interpreted from multi-temporal Sentinel-2 images (Figure 2 and S3 in the Supplemental). Changes through time were inferred by comparing the 2018 coseismic inventory with the post seismic inventories of 2019–2022.

Topographic change detection

Here, we used DEMs with the highest accuracy and selected four topographic factors (relief, slope angle, aspect, and roughness) typically used in landslide susceptibility studies (Dou et al., 2015; Merghadi et al., 2020; Pham et al., 2020; Ren et al., 2014). We used the DEM difference method to track the change in relief, slope, aspect, and surface roughness through time (Ren et al., 2014). The topographic relief, slope angle, aspect, and surface roughness values were then extracted directly from the DEM in ArcGIS 10.3.

Relief represents the actual difference in elevation in a unit area with respect to its local base level, that is, the vertical difference between the highest and the lowest points. We calculated the relief from the DEM by measuring the difference between the minimum and maximum elevations within a 3 × 3 pixels moving window around each pixel in a GIS environment.

Roughness characterizes the variability or irregularity in elevation within a spatial unit and is defined as the ratio of the surface area to the projected area of a given site. We computed the roughness of a pixel using focal statistics of the elevation values within a 3 × 3 pixel neighborhood following Evans (1972).

To facilitate the analysis, the values of each topographic factor were classified into low, medium, and high (Table 2).

Table 2.

Classifications for the topographic analysis.

Type	Relief (m)	Slope angle (°)	Roughness
Low	0-150	0-6	0-1.1
Medium	150-250	6-25	1.1-1.3
High	>250	>25	>1.3

Note: the classification of topographic changes is based on the Geospatial Information Authority of Japan (https://www.gsi.go.jp/kiban/index.html).

3.4 Time-series sentinel-2 analysis

Vegetation phenology can efficiently be obtained at the regional scale from remote sensing images (Song et al., 2017). We used the 10-m Sentinel-2 satellite scenes that cover the whole study area (∼512 km²) to evaluate the vegetation regrowth in the coseismic landslide areas (Table 2). Several studies showed that multitemporal NDVI indicators can be used for vegetation cover classification and to understand vegetation dynamics (Lin et al., 2006; Teillet, 1997; Wu et al., 2004; Yunus et al., 2020). NDVI was found to be a more powerful indicator than the Enhanced Vegetation Index (EVI) in areas with poor vegetation growth (Testa et al., 2014), such as the areas affected by coseismic landslides in our study. The NDVI can be obtained from Sentinel-2 reflectivity band 4 (RED) and band 8 (NIR—Near InfraRed) data by the following formula (Justice et al., 1985):

N D V I = \frac{N I R - R E D}{N I R + R E D}

(1)

Six cloud-free scenes—two pre-earthquake (June 17, 2017, September 5, 2017) and four post-earthquake (September 15, 2018, May 23, 2019, June 1, 2020, June 26, 2021)—were selected for the analysis. We verified that the pre-earthquake data acquired from May to September were comparable: the average NDVI values over the coseismic landslide areas were practically identical (mean difference: 0.0023).

Coseismic landslides mapped by Dou et al. (2020) were used as the vegetation damage zones in this study. Regrowth of vegetation in the coseismic landslide site can be used as a proxy for the lack of significant landslide activity (Yang et al., 2018a; Yunus et al., 2020). The vegetation recovery rate (VRR), calculated from multi-temporal NDVI data, is an effective index to quickly evaluate and monitor the vegetation restoration status within the landslide boundaries (Jiang et al., 2015; Lin et al., 2005). The VRR is calculated as follows:

V R R = \frac{{N D V I}_{1} - {N D V I}_{2}}{{N D V I}_{1} - {N D V I}_{0}} \times 100

(2)

where NDVI₀ is the NDVI value before the earthquake (June 17th, 2017), and NDVI₁ is the NDVI immediately after the earthquake (here, we used Sentinel-2 scene of September 15, 2018), whereas NDVI₂ indicates the NDVI in subsequent scenes. Here, for NDVI₂, we used Sentinel-2 images acquired on May 23, 2019, June 1, 2020, June 26, 2021, and May 17, 2022. The obtained VRR values were tabulated using landslide polygons to assess active and inactive landslides. Based on Lin et al. (2005), VRR values were divided into six classes as listed in Table 3. The classes named “very good” and “extremely good” indicate the best vegetation recovery. Namely, a VRR value above 75 indicates good to excellent vegetation recovery, thereby suggesting absence of landslide activity in the analyzed pixel (Lin et al., 2005, 2006; Yunus et al., 2020). New landslides and remobilized landslides were identified manually from Google Earth, Planet, and Sentinel-2 images by visual interpretation.

Table 3.

Classes of vegetation recovery rate (VRR).

VRR class	VRR value
Extremely poor	<0
Poor	0–25
Medium	25–50
Good	50–75
Very good	75–100
Extremely good	>100

Results

Topographic changes

The study area features a low-elevation hilly terrain (Figure 2(a)). The mean relief evaluated from the pre-earthquake DEM in the coseismic landslide areas is ∼151.87 m, while the mean relief measured after the earthquake is ∼150.11 m (difference: −1.76 m). Following the earthquake, it can be seen that relief in medium- and high-relief areas (150–300 m) has decreased greatly, while it has increased in low-relief areas (0–150 m). In fact, most of the coseismic landslides disrupted ridge crests and the accumulated materials deposited in the valleys (Petle, 2018). It can be seen from Figure 3(e–f) that obvious changes in relief occurred on both ridges (S₁ to S₉) and valleys (V₁ to V₇). Ridges were reduced by as much as 16 m (see S₁), while the elevation of valleys increased by up to 12 m (see V₇). Figure 2(b) shows that the earthquake caused the proportion of slopes facing in E, SE, and S diminishes, whereas W, NW, and N facing slopes gains the spatial domain. This can be related to the orientation of the seismogenic source and its slip (USGS Moment Tensor; see Figure 1), and is generally in agreement with earlier studies (e.g., Fan et al., 2018b). Changes in slope angle due to coseismic landsliding are also evident (Figure 2(c)). In general, earthquake-induced landslides produced a steepening of the landscape as they left steep scarps exposed in their source areas.

Changes in roughness follow a similar pattern (Figure 2(d)), as the proportion of high-roughness areas increased at the expenses of low-roughness areas. We validated these results with 1 m pre- and post-earthquake DEM-derived factors for the 29 km² area and found results in agreement with the 10 m DEM analysis, despite the latter’s lower accuracy and different mode of acquisition (Supplemental Fig. S4).

Time-series analysis of landslides and vegetation pattern

In the Sentinel-2 scene recorded in 2019–2020 (Figure 4(a)), we identified 53 new landslides (0.07 km²) and 29 landslide remobilizations (∼0.1 km²). During 2021–2022, we located further 27 new landslides (0.023 km²) and 31 landslide remobilizations (0.08 km²). Figure 6(a) shows the spatial distribution map of remobilized landslides and new landslides in the first 4 years following the earthquake. Representative examples are shown in Figures 4(b)–(d). The total number of post-seismic new landslides (80) and remobilized landslides (60) is among the lowest in recent strong earthquakes. For example, after the M_w 6.5 Jiuzhaigou earthquake in China, 951 new landslides were identified in the epicentral area with a gross volume of 23.05 × 10⁶ m³ within the first year (Wang and Mao, 2022).

NDVI analysis show that before the earthquake, the coseismic landslide areas were well covered by vegetation (natural and man-made forests and grass fields). The average NDVI value reached 0.87, and about 90% of the area had NDVI values >0.8 (Figure 5). Coseismic landslides affected an area of ∼27 km². In almost all of the area (99%), vegetation has been damaged to varying degrees. After the earthquake, the average NDVI value decreased to 0.36 in the coseismic landslide areas, and its frequency distribution also changed greatly (Figure 5). NDVI values <0.4 were recorded in almost 70% of the areas of coseismic landslides. In some areas, the NDVI in 2019 (1 year after the earthquake) was even lower than that in 2018, with values as low as 0.2. This suggests either that continued landslide activity had damaged the vegetation further or that the damaged vegetation had decayed completely. During 2020–2021 period, the NDVI classes between 0.1 and 0.6 showed a recovery, suggesting an improvement in the prevailing condition and possible slope stabilization. However, because of heavy snowfall in the winter of 2021–2022, the NDVI values found in the 0.6–0.8 and 0.8–1.0 classes decreased to a minimum, which is indicative of decaying vegetation and prevailing soil erosion (Figure 5).

Additionally, we chose the areas that did not suffer from earthquake and compared the NDVI in non-landslide areas. We found that a large proportion of NDVI values fell in the 0.2–0.6 range and the mean NDVI value was ∼0.3, indicating that the earthquake had only a slight impact on vegetation in non-landslide areas (see Supplemental Fig. S5).

VRR results show that in general, a poor vegetation recovery was recorded in coseismic landslide areas (Figures 6 and 7). Areas with high landslide density did not show significant recovery in 2019 (Figure 6(a)), while areas with comparatively low landslide density showed some signs of vegetation recovery (Figure 6(b)). In 2020 and later periods, areas with high landslide density also began to show signs of recovery, albeit only in downslope areas (Figure 6(c), Table 4). Meanwhile, areas with low landslide density achieved good recovery (Figure 6(d)). In general, only 0.50 km² of coseismic landslide areas (i.e., only ∼1%) showed good vegetation recovery in 2019; this area increased to 3.77 km² (∼8.5%) by 2020, and to about 17.82 km² (40.46%) by 2021 (Table 4). However, during 2022, because of heavy snowfall in the winter of 2021–2022, almost all the regrown grasslands were destroyed (Table 4). The slow recovery in large landslides may be attributed to their greater depth and exposed sandy layers devoid of organic matter. Since the natural recovery of vegetation is slow, artificial restoration has been implemented in slope failures on the roadsides for protection. Some examples of vegetation restoration by artificial methods are shown in Figure 8 and Supplemental Fig. S6.

Table 4.

VRR classification statistics.

Year	Extremely poor		Poor		Medium		Good		Very good		Extremely good
Year	Area	%	Area	%	Area	%	Area	%	Area	%	Area	%
2019	24.31	55.21	11.60	26.34	5.55	12.60	2.07	4.70	0.22	0.51	0.28	0.64
2020	22.73	51.61	7.33	16.65	5.35	12.14	4.86	11.03	3.30	7.49	0.47	1.07
2021	7.97	18.10	6.17	14.02	5.54	12.58	6.54	14.85	15.56	35.33	2.26	5.13
2022	34.06	77.33	8.95	20.33	0.48	1.10	0.05	0.12	0.03	0.07	0.46	1.04

Discussion

Peculiar topography of Eastern Iburi

While analyzing global earthquake-induced landslide inventories, Tanyaş et al. (2017) showed that ∼80% of the coseismic landslides occurred in areas with steep slopes and high relief (27° and 524 m, on average, respectively). During the Wenchuan earthquake, coseismic landslides occurred on even steeper slopes, in regions with very high relief (35° and 916 m, on average, respectively) (Ren et al., 2014). In the case of the M_w 6.6 Eastern Iburi earthquake, the numerous coseismic landslides were concentrated on gentle slopes (10–25°, 20° on average). Furthermore, 99% of coseismic landslides occurred at locations where relief was <300 m, and 63% of them occurred at elevations of 100 to 200 m (Wang et al., 2019a). In a similar geoenvironment, for example, during the 2004 Niigata M6.8 earthquake, most of the landslides (>45%) occurred in grid cells with average slopes of 20–35° (Wang et al., 2007). The average relief at coseismic landslide locations in Hokkaido is ∼150 m, comparatively lower than that of the Niigata case (∼250 m). Notably, the slope angle and relief at which landsliding occurred in the Hokkaido event were much gentler than in other recent cases, such as the 1999 Chi-Chi, 2004, Niigata, 2008 Wenchuan, and 2015 Gorkha earthquakes. This peculiar topography, comprising the very weak and slope draping lithology and loose volcanic material, may explain why some global landslide susceptibility models trained in other areas failed to produce a reasonable assessment in the Hokkaido region (e.g., Wang et al., 2019b).

The Eastern Iburi topography can also explain why the topographic changes caused by the earthquake-induced landsliding differ from those observed in other major earthquakes (Figure 5). In fact, our analysis shows an increase in low-relief regions at the expense of high-relief regions and, at the same time, an overall steepening and roughening of the landscape. In the Wenchuan earthquake region, Ren et al. (2014) observed an increase in medium-relief areas and a decreased proportion of steep slopes. Their study pointed out that after the earthquake, the topography in the Longmen Shan became smoother. The increased angle of slope and roughness in Eastern Iburi may suggest relatively less stable topographic conditions, which is concerning.

Furthermore, the recovery rate of vegetation in the earthquake-affected region is related to the size of landslide areas. We compared the size-frequency distribution of coseismic landslides of the Iburi earthquake with that of other cases worldwide in Supplemental Fig. S7, which shows that the Iburi earthquake triggered landslides of an average size among other recent earthquakes. Furthermore, several new landslides occurred in 2019 and 2020 after the Iburi earthquake, whereby their size-frequency had a similar trend to that seen after other earthquakes.

Evolution of post-seismic landslides

We observed only a small number of landslide remobilizations and new landslides in the study area after the earthquake (Figure 4). One possible reason is that the debris comprising volcanic ash and pumice materials mobilized through the coseismic landsliding accumulated quickly in the valley bottoms (Li et al., 2018a, 2018b) because of the smaller relief in the region. Conversely, in other earthquake-stricken regions, the availability of erodible materials on hillslopes was associated with the more frequent remobilizations of coseismic landslide deposits evolving into large debris flows (Dahlquist and West 2019; Tang et al., 2009; Zhang and Zhang 2017). In some cases, such as in the Chi-Chi and Jiuzhaigou areas, mostly new landslides were reported in the initial years after the earthquake (Fan et al., 2019; Hovius et al., 2011). For the Wenchuan and Gorkha cases, most of the materials remain on the hillslope, and their evacuation via landsliding and other erosional processes may take a long time (Yunus et al., 2020).

In Eastern Iburi, we hypothesize three reasons for the comparatively low landslide activity after the earthquake: (i) low amount of rainfall and snowfall during the initial years following the earthquake; (ii) limited accumulation of sediments on hillslopes; and (iii) limited damage to the bedrock (because of the comparatively small earthquake magnitude) that could have triggered new landslides over time, while coseismic landslides mostly affected the pyroclastic materials covering the slopes (3–5 m thick) (Kawamura et al., 2019). A schematic of the geological profile before and after the earthquake is given in Fig. S8. Note the sandy nature of Shikotsu pumice falls in Fig. S8, that might have contributed to the low vegetation recovery in places having deeper landslide depths. Nevertheless, while (i) and (ii) are based on actual observations, (iii) remains speculative as only time and future wetter/warmer seasons may clarify how much of the slope and bedrock was weakened by the earthquake.

Vegetation as a proxy of landslide stabilization

The recovery of vegetation in coseismic landslide areas can reflect the progressive decrease in landslide activity and inform on the stability of the slopes (Korup et al., 2007; Mondini et al., 2011; Yang et al., 2018b). However, some studies have recognized a lag effect between landslide activity and vegetation recovery that depends on the progressive development of vegetation roots and soil consolidation (Chen et al., 2020). After the Wenchuan earthquake, vegetation recovered slowly (∼10 years to achieve an 83% recovery; Yunus et al., 2020), yet the landslide activity remained elevated (Chen et al., 2020; Huang and Li 2014). In the Niigata–Chuetsu area, ∼88% recovery was achieved in less than 5 years (Yunus et al., 2020) and landslide activity largely decreased during the same period (Marc et al., 2015). In our study area, a good vegetation regrowth was only observed in areas with low landslide density. Conversely, areas with high coseismic landslide density, having been more severely damaged, did not show much regrowth in the first 4 years after the earthquake. This may be the result of coseismic landslides fully removing the humic soil cover and exposing the sandy layer of volcanic material or the underlying bedrock but can also relate to the poor stability and ongoing deformation of coseismic landslide bodies, even in absence of evident remobilizations. Continued monitoring of vegetation regrowth remains critical to understand the possibility of future disasters. Based on the initial rate of vegetation recovery and the trends observed in similar regions (Yunus et al., 2020), it could take about 15 years for vegetation to recover to the pre-earthquake level, and even longer for the landslide areas to fully stabilize. However, it is expected that extreme rainfall or snowmelt events may substantially alter the current trend (e.g., Shou et al., 2011a; Shou et al., 2011b).

The results discussed in this study suggest the possibility of ongoing post-seismic debris activities and gully development in the region, emphasizing the need for continued assessment. In several earthquake-affected sites, it has been noticed that post-seismic landslide activity escalated in the initial years. We therefore recommend continued slope monitoring until stabilization of vegetation recovery. Crucial indicators of ecological restoration other than NDVI, such as leaf area index (LAI), enhanced vegetation index (EVI), and net primary productivity (NPP) may be combined for a complete understanding of the ecological restoration in the region. As SAR data provide cloud-free solutions, future SAR acquisition data are poised to provide crucial technical support for accurately mapping the vegetation trends.

Conclusion

The 2018 M_w 6.6 Eastern Iburi Earthquake triggered over 8500 landslides in Hokkaido’s Atsuma region, offering a unique opportunity to assess post-seismic surface shifts and their impact on vegetation. We utilized high-resolution pre- and post-event topographic data to analyze the dynamics and effects of the seismic event. Our study tracked topographic alterations and vegetation responses through multitemporal remote sensing data. Contrary to expectations, the initially gentle terrain in Eastern Iburi became rougher, with increased slope steepness post-event. The limited occurrence of new landslides and remobilizations suggests ample debris remains on hillslopes and in valleys owing to below-average rainfall and snowfall. Nevertheless, ongoing monitoring is essential to gauge the effects of precipitation in the coming years, as studies indicate that an elevated likelihood of landslides can persist for years following a significant earthquake. Thanks to the enhanced DEM resolution, the conclusions drawn from our study are characterized by reliability and have low uncertainty. This not only strengthens the validity of our findings but also makes the data usable in further studies (e.g., about erosion calculations, landscape evolution). Furthermore, our observations reveal poor overall vegetation recovery in the study area, with the NDVI value dropping from 0.86 to 0.36 in coseismic landslide zones. Over the first 4 years post-earthquake, vegetation recovery rates in zones categorized as good to very good were notably low, falling below 40%. Where recovery is evident, it primarily occurs in small landslides and consists mainly of grasslands and artificially restored slopes.

Supplemental Material

Supplemental Material - Post-seismic topographic shifts and delayed vegetation recovery in the epicentral area of the 2018 Mw 6.6 Hokkaido Eastern Iburi earthquake

Supplemental Material for Post-seismic topographic shifts and delayed vegetation recovery in the epicentral area of the 2018 Mw 6.6 Hokkaido Eastern Iburi earthquake by Jie Dou, Zilin Xiang, Xiekang Wang, Penglin Zheng, Ram Avtar, Chen Xinyu, Gianvito Scaringi, Wanqi Luo, and Ali P Yunus in Progress in Physical Geography: Earth and Environment.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China; 42090054; 51639007.

ORCID iDs

Dou Jie

Xiekang Wang

Ali P Yunus

Data availability statement

The pre-earthquake 10 m digital elevation model used in this study for topographic analysis are freely available from Geospatial Information Authority of Japan (GSI: https://www.gsi.go.jp/). Pre- and post-earthquake 1 m digital elevation model is obtained from the Local Hokkaido Government (https://www.geospatial.jp/). Coseismic landslide boundary shapefiles are available to download subject to the publication in an open repository (https://10.5281/zenodo.5618609). Post-seismic landslides are mapped from Sentinel-2 images and are downloaded from .

Supplemental Material

Supplemental material for this article is available online.

References

Asano

Iwata

(2019) Source rupture process of the 2018 Hokkaido Eastern Iburi earthquake deduced from strong-motion data considering seismic wave propagation in three-dimensional velocity structure. Earth Planets and Space 71: 101. DOI: 10.1186/s40623-019-1080-0.

Bontemps

Lacroix

Larose

, et al. (2020) Rain and small earthquakes maintain a slow-moving landslide in a persistent critical state. Nature Communications 11: 780. DOI: 10.1038/s41467-020-14445-3.

Chen

R-F

Chang

K-J

Angelier

, et al. (2006) Topographical changes revealed by high-resolution airborne LiDAR data: the 1999 Tsaoling landslide induced by the Chi–Chi earthquake. Engineering Geology 88: 160–172. DOI: 10.1016/j.enggeo.2006.09.008.

Chen

Tang

Xiong

, et al. (2020) The long-term evolution of landslide activity near the epicentral area of the 2008 Wenchuan earthquake in China. Geomorphology 367: 107317. DOI: 10.1016/j.geomorph.2020.107317.

Chigira

Yagi

(2006) Geological and geomorphological characteristics of landslides triggered by the 2004 Mid Niigta prefecture earthquake in Japan. Engineering Geology 82: 202–221. DOI: 10.1016/j.enggeo.2005.10.006.

Croissant

Lague

Steer

, et al. (2017) Rapid post-seismic landslide evacuation boosted by dynamic river width. Nature Geoscience 10: 680–684. DOI: 10.1038/ngeo3005.

Dahlquist

West

(2019) Initiation and runout of post-seismic debris flows: insights from the 2015 Gorkha earthquake. Geophysical Research Letters 46: 9658–9668. DOI: 10.1029/2019GL083548.

Dai

Scaringi

Fan

, et al. (2021) Coseismic debris remains in the orogen despite a decade of enhanced landsliding. Geophysical Research Letters 48: e2021GL095850. DOI: 10.1029/2021GL095850.

Domènech

Fan

Scaringi

, et al. (2019) Modelling the role of material depletion, grain coarsening and revegetation in debris flow occurrences after the 2008 Wenchuan earthquake. Engineering Geology 250: 34–44. DOI: 10.1016/j.enggeo.2019.01.010.

10.

Dou

Chang

Chen

, et al. (2015) Automatic case-based reasoning approach for landslide detection: integration of object-oriented image analysis and a genetic algorithm. Remote Sensing 7: 4318–4342. DOI: 10.3390/rs70404318.

11.

Dou

Yunus

Merghadi

, et al. (2020) Different sampling strategies for predicting landslide susceptibilities are deemed less consequential with deep learning. The Science of the Total Environment 720: 137320. DOI: 10.1016/j.scitotenv.2020.137320.

12.

Fan

Domènech

Scaringi

, et al. (2018a) Spatio-temporal evolution of mass wasting after the 2008 Mw 7.9 Wenchuan earthquake revealed by a detailed multi-temporal inventory. Landslides 15: 2325–2341. DOI: 10.1007/s10346-018-1054-5.

13.

Fan

Scaringi

, et al. (2018b) Coseismic landslides triggered by the 8th August 2017 Ms 7.0 Jiuzhaigou earthquake (Sichuan, China): factors controlling their spatial distribution and implications for the seismogenic blind fault identification. Landslides 15: 967–983. DOI: 10.1007/s10346-018-0960-x.

14.

Fan

Scaringi

(2018c) Brief communication: post-seismic landslides, the tough lesson of a catastrophe. Natural Hazards and Earth System Sciences 18: 397–403. DOI: 10.5194/nhess-18-397-2018.

15.

Fan

Scaringi

Korup

, et al. (2019) Earthquake-induced chains of geologic hazards: patterns, mechanisms, and impacts. Reviews of Geophysics 57: 421–503. DOI: 10.1029/2018RG000626.

16.

Fan

Yunus

Scaringi

, et al. (2021) Rapidly evolving controls of landslides after a strong earthquake and implications for hazard assessments. Geophysical Research Letters 48(1): e2020GL090509.

17.

Gan

Yang

Zhang

, et al. (2019) Temporal and spatial evolution of vegetation coverage in the mianyuan river basin influenced by strong earthquake disturbance. Scientific Reports 9(1): 16762.

18.

Hovius

Meunier

Lin

, et al. (2011) Prolonged seismically induced erosion and the mass balance of a large earthquake. Earth and Planetary Science Letters 304: 347–355. DOI: 10.1016/j.epsl.2011.02.005.

19.

Hua

Zhao

, et al. (2019) Arc-Arc collision caused the 2018 eastern Iburi earthquake (M 6.7) in Hokkaido, Japan. Scientific Reports 9: 13914–13919. DOI: 10.1038/s41598-019-50305-x.

20.

Huang

(2014) Post-earthquake landsliding and long-term impacts in the Wenchuan earthquake area, China. Engineering Geology 182: 111–120. DOI: 10.1016/j.enggeo.2014.07.008.

21.

Jiang

W-G

Jia

J-J

, et al. (2015) Evaluating the vegetation recovery in the damage area of wenchuan earthquake using MODIS data. Remote Sensing 7: 8757–8778. DOI: 10.3390/rs70708757.

22.

JMA (2018). The 2018 Hokkaido Eastern Iburi earthquake -portal- . Japan Meteorological Agency. https://www.jma.go.jp/jma/en/2018_Hokkaido_Eastern_Iburi_Earthquake/index.html

23.

Justice

Townshend

JRG

Holben

, et al. (1985) Analysis of the phenology of global vegetation using meteorological satellite data. International Journal of Remote Sensing 6: 1271–1318. DOI: 10.1080/01431168508948281.

24.

Kawamura

Kawajiri

Hirose

, et al. (2019) Slope failures/landslides over a wide area in the 2018 Hokkaido Eastern Iburi earthquake. Soils and Foundations 59: 2376–2395. DOI: 10.1016/j.sandf.2019.08.009.

25.

Keefer

(1994) The importance of earthquake-induced landslides to long-term slope erosion and slope-failure hazards in seismically active regions. In: Geomorphology and Natural Hazards. Amsterdam, Netherlands: Elsevier, 265–284.

26.

Kita

(2019) Characteristics of relocated hypocenters of the 2018 M6.7 Hokkaido Eastern Iburi earthquake and its aftershocks with a three-dimensional seismic velocity structure. Earth Planets and Space 71: 122. DOI: 10.1186/s40623-019-1100-0.

27.

Kobayashi

Koketsu

Miyake

(2019) Rupture process of the 2018 Hokkaido Eastern Iburi earthquake derived from strong motion and geodetic data. Earth Planets and Space 71: 63. DOI: 10.1186/s40623-019-1041-7.

28.

Kober

Zeilinger

Hippe

, et al. (2015) Tectonic and lithological controls on denudation rates in the central Bolivian Andes. Tectonophysics 657: 230–244.

29.

Korup

Clague

Hermanns

, et al. (2007) Giant landslides, topography, and erosion. Earth and Planetary Science Letters 261: 578–589. DOI: 10.1016/j.epsl.2007.07.025.

30.

West

Densmore

, et al. (2017) Earthquakes drive focused denudation along a tectonically active mountain front. Earth and Planetary Science Letters 472: 253–265.

31.

Wang

Liu

(2018a) A decadal evolution of landslides and debris flows after the Wenchuan earthquake. Geomorphology 323: 1–12. DOI: 10.1016/j.geomorph.2018.09.010.

32.

Wang

Liu

, et al. (2018b) Topographic changes and their driving factors after 2008 Wenchuan earthquake. Geomorphology 311: 27–36. DOI: 10.1016/j.geomorph.2018.03.019.

33.

Lin

W-T

Chou

W-C

Lin

C-Y

, et al. (2005) Vegetation recovery monitoring and assessment at landslides caused by earthquake in Central Taiwan. Forest Ecology and Management 210: 55–66. DOI: 10.1016/j.foreco.2005.02.026.

34.

Lin

W-T

Lin

C-Y

Chou

W-C

(2006) Assessment of vegetation recovery and soil erosion at landslides caused by a catastrophic earthquake: a case study in Central Taiwan. Ecological Engineering 28: 79–89. DOI: 10.1016/j.ecoleng.2006.04.005.

35.

Marc

Hovius

Meunier

, et al. (2015) Transient changes of landslide rates after earthquakes. Geology 43(1): 883–886. DOI: 10.1130/G36961.

36.

Massey

McSaveney

Taig

, et al. (2014) Determining rockfall risk in Christchurch using rockfalls triggered by the 2010-2011 Canterbury earthquake sequence. Earthquake Spectra 30(1): 155–181. DOI: 10.1193/021413EQS026M.

37.

Merghadi

Yunus

Dou

, et al. (2020) Machine learning methods for landslide susceptibility studies: a comparative overview of algorithm performance. Earth-Science Reviews 207: 103225. DOI: 10.1016/j.earscirev.2020.103225.

38.

Meunier

Hovius

Haines

(2007) Regional patterns of earthquake-triggered landslides and their relation to ground motion. Geophysical Research Letters 34: L20408. DOI: 10.1029/2007GL031337.

39.

Mondini

Guzzetti

Reichenbach

, et al. (2011) Semi-automatic recognition and mapping of rainfall induced shallow landslides using optical satellite images. Remote Sensing of Environment 115: 1743–1757. DOI: 10.1016/j.rse.2011.03.006.

40.

Nam

Wang

(2019) The performance of using an autoencoder for prediction and susceptibility assessment of landslides: a case study on landslides triggered by the 2018 Hokkaido Eastern Iburi earthquake in Japan. Geoenvironmental Disasters 6: 19. DOI: 10.1186/s40677-019-0137-5.

41.

Parker

Hancox

Petley

, et al. (2015) Spatial distributions of earthquake-induced landslides and hillslope preconditioning in the northwest South Island, New Zealand. Earth Surface Dynamics 3: 501–525. DOI: 10.5194/esurf-3-501-2015.

42.

Petle

(2018) Landslides Triggered by the 6th September 2018 Hokkaido Earthquake. EOS, The Landslide Blog. Available at: https://blogs.agu.org/landslideblog/2018/09/06/landslides-6th-september-2018-hokkaido-earthquake/.

43.

Pham

Prakash

Dou

, et al. (2020) A novel hybrid approach of landslide susceptibility modelling using rotation forest ensemble and different base classifiers. Geocarto International 35: 1267–1292. DOI: 10.1080/10106049.2018.1559885.

44.

Ren

Zhang

Dai

, et al. (2014) Topographic changes due to the 2008 Mw7.9 Wenchuan earthquake as revealed by the differential DEM method. Geomorphology 217: 122–130. DOI: 10.1016/j.geomorph.2014.04.020.

45.

Roback

Clark

West

, et al. (2018) The size, distribution, and mobility of landslides caused by the 2015 Mw7. 8 Gorkha earthquake, Nepal. Geomorphology 301: 121–138.

46.

Saba

van der Meijde

van der Werff

(2010) Spatiotemporal landslide detection for the 2005 Kashmir earthquake region. Geomorphology 124: 17–25. DOI: 10.1016/j.geomorph.2010.07.026.

47.

Shao

, et al. (2019) Planet image-based inventorying and machine learning-based susceptibility mapping for the landslides triggered by the 2018 Mw6. 6 Tomakomai, Japan Earthquake. Remote Sensing 11(8): 978.

48.

Shen

Zhang

Fan

, et al. (2020) Declining geohazard activity with vegetation recovery during first ten years after the 2008 Wenchuan earthquake. Geomorphology 352: 106989. DOI: 10.1016/j.geomorph.2019.106989.

49.

Shou

Hong

, et al. (2011a) Spatial and temporal analysis of landslides in Central Taiwan after 1999 Chi-Chi earthquake. Engineering Geology 123: 122–128. DOI: 10.1016/j.enggeo.2011.03.014.

50.

Shou

Fei

, et al. (2011b) Dynamic environment in the Ta-chia river watershed after the 1999 taiwan chi-chi earthquake. Geomorphology 133: 190–198. DOI: 10.1016/j.geomorph.2011.04.038.

51.

Song

Ruan

, et al. (2017) Estimating fractional vegetation cover and the vegetation index of bare soil and highly dense vegetation with a physically based method. International Journal of Applied Earth Observation and Geoinformation 58: 168–176. DOI: 10.1016/j.jag.2017.01.015.

52.

Steer

Jeandet

Cubas

, et al. (2020) Earthquake statistics changed by typhoon-driven erosion. Scientific Reports 10: 10899. DOI: 10.1038/s41598-020-67865-y.

53.

Takahashi

Takai

Chigira

, et al. (2020) Special issue “the 2018 Hokkaido eastern Iburi earthquake and hidaka arc–arc collision system.”. Earth Planets and Space 72: 65–67.

54.

Tang

Zhu

, et al. (2009) Rainfall-triggered debris flows following the Wenchuan earthquake. Bulletin of Engineering Geology and the Environment 68: 187–194. DOI: 10.1007/s10064-009-0201-6.

55.

Tang

Tanyas

van Westen

, et al. (2019) Analysing post-earthquake mass movement volume dynamics with multi-source DEMs. Engineering Geology 248: 89–101. DOI: 10.1016/j.enggeo.2018.11.010.

56.

Tanyaş

van Westen

Allstadt

, et al. (2017) Presentation and analysis of a worldwide database of earthquake-induced landslide inventories. Journal of Geophysical Research: Earth Surface 122: 1991–2015. DOI: 10.1002/2017JF004236.

57.

Teillet

(1997) Effects of spectral, spatial, and radiometric characteristics on remote sensing vegetation indices of forested regions. Remote Sensing of Environment 61: 139–149. DOI: 10.1016/S0034-4257(96)00248-9.

58.

Testa

Mondino

ECB

Pedroli

(2014) Correcting MODIS 16-day composite NDVI time-series with actual acquisition dates. European Journal of Remote Sensing 47: 285–305. DOI: 10.5721/EuJRS20144718.

59.

Wang

Mao

(2022) Spatio-temporal evolution of post-seismic landslides and debris flows: 2017 M s 7.0 Jiuzhaigou earthquake. Environmental Science and Pollution Research 29(11): 15681–15702.

60.

Wang

Sassa

(2007) Analysis of a spatial distribution of landslides triggered by the 2004 Chuetsu earthquakes of Niigata Prefecture, Japan. Natural Hazards 41: 43–60. DOI: 10.1007/s11069-006-9009-x.

61.

Wang

Fan

Yunus

, et al. (2019a) Coseismic landslides triggered by the 2018 Hokkaido, Japan (Mw 6.6), earthquake: spatial distribution, controlling factors, and possible failure mechanism. Landslides 16: 1551–1566. DOI: 10.1007/s10346-019-01187-7.

62.

Wang

Fan

Yunus

, et al. (2019b) Coseismic landslides triggered by the 2018 Hokkaido, Japan (Mw 6.6), earthquake: spatial distribution, controlling factors, and possible failure mechanism. Landslides 16: 1551–1566.

63.

Yan

, et al. (2004) Developing method of vegetation fraction. IEEE 6: 4352.

64.

Xiang

Dou

Yunus

, et al. (2023) Vegetation-landslide nexus and topographic changes post the 2004 Mw 6.6 Chuetsu earthquake. Catena 223: 106946.

65.

Yao

, et al. (2014) Three (nearly) complete inventories of landslides triggered by the May 12, 2008 Wenchuan Mw 7.9 earthquake of China and their spatial distribution statistical analysis. Landslides 11: 441–461.

66.

Yamagishi

Yamazaki

(2018) Landslides by the 2018 Hokkaido iburi-tobu earthquake on september 6. Landslides 15: 2521–2524. DOI: 10.1007/s10346-018-1092-z.

67.

Yang

Zhou

(2018a) Effects of precipitation and topography on vegetation recovery at landslide sites after the 2008 Wenchuan earthquake. Land Degradation & Development 29: 3355–3365. DOI: 10.1002/ldr.3098.

68.

Yang

Zhou

(2018b) Decreased post-seismic landslides linked to vegetation recovery after the 2008 Wenchuan earthquake. Ecological Indicators 89: 438–444. DOI: 10.1016/j.ecolind.2017.12.006.

69.

Yunus

Fan

Tang

, et al. (2020) Decadal vegetation succession from MODIS reveals the spatio-temporal evolution of post-seismic landsliding after the 2008 Wenchuan earthquake. Remote Sensing of Environment 236: 111476. DOI: 10.1016/j.rse.2019.111476.

70.

Yunus

Xinyu

Catani

, et al. (2023) Earthquake-induced soil landslides: volume estimates and uncertainties with the existing scaling exponents. Scientific Reports 13(1): 8151.

71.

Zang

Shen

(2019) Rupture directivity analysis of the 2018 Hokkaido eastern Iburi earthquake and its seismotectonic implication. Seismological Research Letters 90: 2121–2131. DOI: 10.1785/0220190131.

72.

Zhang

(2017) Impact of the 2008 Wenchuan earthquake in China on subsequent long-term debris flow activities in the epicentral area. Geomorphology 276: 86–103. DOI: 10.1016/j.geomorph.2016.10.009.

73.

Zhang

Wang

, et al. (2019) Characteristics of landslides triggered by the 2018 Hokkaido eastern Iburi earthquake, northern Japan. Landslides 16: 1691–1708. DOI: 10.1007/s10346-019-01207-6.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.51 MB