Inventory of shallow landslides triggered by extreme precipitation in July 2023 in Beijing, China

Inventory of shallow landslides triggered by extreme precipitation in July 2023 in Beijing, China | Scientific Data

Abstract

The extreme meteorological events caused by climate change have increasingly caused serious clustered landslides. A systematic and timely inventory of triggered landslides is a prerequisite for quantifying and evaluating the impact of extreme meteorological events. In addition, a landslide inventory can provide basic data for any subsequent analysis of the event or be used innovatively in landslide risk analysis. The Rainfall-induced Landslides in Beijing (RLBJ) inventory presented here contains data on 15,383 rainfall-induced shallow landslides triggered by a single extreme precipitation event in July 2023 in an area of ~3,250 km2 in western mountainous areas of Beijing, China. High-resolution satellite images before and after this rainstorm event were used to visually analyze the landslides. All landslides were reported as vectorized points and polygon features and classified according to their motion forms. This inventory is now freely available for the benefit of international geohazard researchers.

Background & Summary

Landslides and related mass movement processes are widespread natural hazards that occur in mountainous areas and frequently cause severe damage to infrastructure, properties, and human life 1, 2, 3, 4. Large or clustered landslides can also significantly affect the landscape of an area 5, 6, 7, 8, 9. Rainfall is a primary trigger of landslides, especially shallow ones 10. Heavy rainstorm events can trigger up to tens of thousands of landslides simultaneously in rainfall-affected regions 11, 12, 13. Against the backdrop of global climate change, extreme precipitation events are becoming increasingly common and are hitting various parts of the world 14, 15, 16, even in many traditionally arid and low-rainfall areas 17. These rare extreme precipitation events in typically less rainy areas are likelier to trigger unexpected landslides 18, 19, 20.

The preparation of landslide maps to register the extent of historical landslides in an area is important for investigating the distribution, types, pattern, and recurrence of slope failures from the slope or catchment scale to regional or national scales 21, 22, 23, 24, 25; assessing landslide susceptibility, hazard, vulnerability and risk 1, 26, 27, 28, 29, 30; and studying the evolution of local or regional landscapes and ecological environment dominated by mass-wasting processes 31, 32, 33, 34, 35, 36. Landslide inventory maps commonly consist of locations and, when available, detailed information on geometry, dates of occurrence, material involved, mass motion forms, damage, and remediation work 13, 25, 37, 38. Vectorized geometric shapes (i.e., polygonal shapes and line or point features) depicted in the geographical information system (GIS) environment are the most common and convenient reporting forms for landslide maps 13, 25, 37, 39. Producing event inventory maps is an important aspect of landslide mapping. An event landslide map can fully describe the extent and impact of landslides after every triggering event (e.g., rainstorms 11, 40, rapid snowmelt events 41, or earthquakes) 39, 42 and contain crucial data for any subsequent analyses of the ground damage effects of the event 13. The information in these maps can also provide valuable guidance for emergency and post-event recovery efforts 36, 38.

Several methods have been developed to map landslides, including field surveys 13, 43; expert interpretations of aerial photographs 44 and optical 45, 46, multispectral 47, or synthetic aperture radar (SAR) satellite images 48; morphometric analyses of digital elevation models (DEMs) 37; etc. Classical field surveys can provide the most detailed information on the landslide masses (e.g., materials, motion path, and deposit area) and is least affected by shading of tree canopies, clouds, or structures; however, it requires significant manpower, material resources, and time and relies on the accessibility of hazards 13. Interpretation based on optical satellite or aerial images can provide a timely approach to the hazard as it is not limited by road blockage; however, cloud cover, a lack of natural lighting, and image resolution can strongly affect the accuracy and efficiency of the interpretation 38. Manual visual identification is the most frequently used method 45, 46, 49. Innovative automatic or semi-automatic algorithms have been developed to improve the efficiency of interpretation; however, their results need to be verified based on data from visual interpretation 50, 51. The SAR data can be used to construct a time series of surface deformations and are suitable for identifying inapparent landslides that are still in the creep stage 48. The DEM contains rich terrain information and is mostly used to assist in optical image interpretation 38, but it can also be used separately to identify large ancient or subaqueous landslides 37, which have clear morphometric features. Therefore, visual interpretation from optical images is still the most feasible way to map event-triggered landslides in a timely and accurate manner.

From July 29, 2023 to August 2, 2023, affected by the residual circulation of Super Typhoon Doksuri 52, a rare extreme precipitation event occurred in Beijing, which is located in an area of northern China that typically does not receive much rainfall. The average precipitation in the city reached 331 mm over 83 hours, or approximately 60% of the city’s annual average precipitation 1)—this resulted in widespread mountain floods and landslides. This event caused a total of 1.29 million people to be affected by the disaster, and 33 human deaths and resulted in 18 persons being declared missing. At least 59,000 houses collapsed, and 147,000 houses were severely damaged 52. This paper describes the processing procedures and record structure of the landslide inventory map for this rare extreme precipitation event. The RLBJ presented here is the result of an extensive remote sensing interpretation for the area most affected by the intense rainfall, spanning ~3250 km2 (Fig. 1).

Location map of the area of interest. Note: The contour cloud map illustrates the spatial arrangement of cumulative precipitation, with data collected from 20:00, July 29, 2023 to 13:00, July 31, 2023 (Source: Beijing Meteorological Service; https://weibo.com/2611704935/NcoOkkeyG?pagetype=profilefeed). Base map: ALOS 12.5 m DEM 53.

Methods

Figure 2 illustrates the methodology workflow adopted to prepare the RLBJ inventory in detail. The identification and delimitation of the landslides was based on the changes detected from optical satellite images taken before (e.g., Fig. 3a) and after (e.g., Fig. 3b) the precipitation event. Three main stages were set for the preparation process: (i) data acquisition, which mainly involves the available optical satellite images; (ii) manually digitizing landslide features; and (iii) inventory map integration and verification.

Flow chart of the procedure used to prepare the RLBJ inventory presented in this paper.

Excerpts of the satellite images used to prepare the inventory and the resulting landslide map. (a) Satellite image before the studied extreme precipitation event. No landslides were visible on the map at this time. (b) Satellite image after the studied extreme precipitation event. The images taken at this time depict massive newly occurring landslides. (c) Landslide map after the interpretation process. Note: All landslides are mapped as points and polygons. The points are marked at the initial positions of the landslides, and the polygons cover the entire affected areas of the landslides. Base map: ALOS 12.5 m DEM 53 and derived contours at 20 m equidistance.

Available data

The data used to produce and verify this inventory mainly include optical satellite images taken before and after the focused rainfall event. Acquisition and viewing of satellite images are fully conducted through the OvitalMap map browser (Version: V10.0.3.32078; Beijing Ovital Software Co., Ltd., https://www.ovital.com), which integrates electronic maps and satellite images from multiple sources. The satellite images used in this study are provided by SIWEIearth ( https://siweiearth.com), and the images can be directly viewed in the OvitalMap map browser. These images have been precisely merged with an overall cloud cover of less than 20% and possess a high resolution (ranging from 0.8 m to 2 m). To target our research areas of interest, we mainly utilized images taken on May 13, 2023 (before the rainfall event; Fig. 3a) and August 19, 2023 (after the rainfall event; Fig. 3b). However, a small portion of the study area is affected by cloud cover and terrain shadows; when this occurred, images from dates before March 15, 2024 were also used. These high-resolution images allowed us to identify the vast majority of landslides triggered by the studied event. The smallest landslide identified from the satellite images had an area of less than 100 m2.

To focus on the most severely affected areas during this extreme precipitation event as the primary study areas, we firstly collected rainfall data (Fig. 1; Source: Beijing Meteorological Service; https://weibo.com/2611704935/NcoOkkeyG?pagetype=profilefeed), news reports 52, ALOS 12.5 m DEM 53, and administrative divisions (Source: CnOpenData; https://www.cnopendata.com/data/m/recent/chinese-administrative-divisio…) as the auxiliary data, and then conducted preliminary field investigations. All of these preliminary efforts indicate that the western mountainous areas of Beijing, which include all mountainous areas within the Fangshan and Mentougou Districts, and the mountainous areas in the southwest of Changping District and northwest of Haidian District, experienced the highest rainfall (Fig. 1) and suffered the most severe property damage caused by the rainfall-induced geological disasters. Therefore, mountainous western areas were designated as our areas of interest, as shown in Fig.

Landslide morphometric mapping

The delimiting and mapping procedure for the landslide areas was carried out using the OvitalMap map browser ( https://www.ovital.com), which provides both 2D and 3D visualization perspectives of the required optical satellite images. The 3D view allows for fully flexible navigation over a virtual mosaic of periodically updated satellite images, including features such as free zoom, rotation, and angle of view. Furthermore, the OvitalMap supports commonly used graphic vectorization operations, including marking points, lines, and polygons. Therefore, the identification, delimiting, and labeling of landslides can be completed in one sitting through the OvitalMap. All identified landslides were marked as points and polygons (Fig. 3c). Prior to mapping the landslide features, we found through our pre-analysis that the landslides triggered by this precipitation event were commonly small slides at first that eventually caused larger disasters by subsequent fluidized motion and entrainment of materials along the motion path. Therefore, all points were located at the initial position of the registered landslide (Fig. 39.

Classifying the registered landslides is another fundamental work for preparing the inventory 38. Classifications according to the material, motion form, and velocity of landslide mass, as proposed by Hungr et al. (2014), are commonly made worldwide 54. However, fully identifying the mass materials and motion velocity through satellite images is infeasible due to the small sizes of many of the studied landslides, despite the images used being of the highest available quality. Therefore, we mainly adopted the motion form to ease and homogenize landslide classification when preparing the RLBJ inventory (Table 1). Table 2 presents typical excerpts of the identification characteristics of each type of landslide. All identifications for the landslide type were based on a 3D view of the images.

Data merging

The manually depicted points and polygons in OvitalMap need to be converted into shapefiles for further merging into the final inventory map in the ArcGIS platform. The final form of the RLBJ inventory is a vectorized multilayer electronic map merged in ArcGIS 10.8 (Table 3). OvitalMap can directly output landslide features (i.e., points and polygons) to the desired shapefile format; however, some post-processing is required for the final map. First, to improve versatility, the map coordinate system was converted to the internationally recognized World Geodetic System-1984 (WGS-84) coordinate system. A topology check was then conducted in the ArcGIS platform to ensure that the points and polygons were matched and to avoid overlapping polygons and other artificial graphical errors. The points and polygons corresponding to the same landslide were assigned a unique ID (UID) to homogenize landslide numbering.

To further enrich the inventory data, some fundamental geometric indices of the landslide features (i.e., area and perimeter) and topographic indices of the initial position (i.e., elevation, topographic position index 55, slope aspect, and slope gradient) were extracted from the ALOS 12.5 m DEM 53 and listed in the attribute tables. The area and perimeter were calculated in the WGS-1984 UTM Zone 50 N projected coordinate system.

Data Records

The RLBJ inventory 56, publicly available for download from Figshare, includes 15,383 total landslide records distributed over the mountainous western areas of Beijing. This inventory can provide an important source for analyzing rainfall-induced landslides in GIS environments.

The dataset consists of a set of shapefiles describing the initial position (point) and area (polygon) features of the landslides and relevant auxiliary geographic elements (Table 3). Table 4 reports the details of the elements included in the attribute tables associated with all feature layers in the RLBJ inventory map.

The spatial distribution and count statistics of the classified landslides are shown in Fig. 4. Of the identified landslides, 98% exhibit different degrees of fluidization mobility. A total of 15,383 landslides are distributed in an area of 3,250 km2, indicating an average density of approximately 4.7 landslides per square kilometer. Figure 5 provides preliminary statistics on landslide point density. The distribution of landslides presents two main concentrated areas in the southeast and northwest of Mentougou District. The maximum point density of landslides in the study area reached 120 per square kilometer.

Location of landslides classified based on movement type. Note: The ring chart shows the number and proportion of each type of landslide. Base map: ALOS 12.5 m DEM 53. The map is based on the WGS-1984 geographic coordinate system.

Point density map of landslides. Note: The horizontal and vertical spacing of the statistical grid is 1 km. The map is projected in the WGS-1984 UTM Zone 50 N coordinate system.

Table 5 summarizes the number and area (AL) of landslides according to the subdivided forms of movement described in Table 1. Overall, the total area of all identified landslides is 19.3 km2 (Table 5), which accounts for 0.59% of the area of interest. An individual landslide ranges ~20 m2 < AL < 1.2 × 105 m2 in the landslide area (AL in m2). Most landslides have an area smaller than 1,000 square meters, accounting for over 70% (Fig. 6a). There are high power–law correlations between the cumulative number of landslides and the landslide area (Fig. 6b).

Statistical charts of landslide area. (a) Histogram of landslide area. (b) Relationship of cumulative number of landslides and landslide area. Note: The N(AL) refers to the number of landslides with an area greater than AL.

Figure 7 summarizes four key topographic factors that may affect the formation of landslides. Landslides are more likely to occur on slopes at elevations between 200 and 800m, particularly on ridges, middle slopes, or valleys, with slope aspects ranging from east to south and gradients between 20°and 35°.

Summary of observations of the initial landslide sites using four basic topographic parameters. (a) Elevation. (b) Topographic position. (c) Slope aspect. (d) Slope gradient. Note: RD—Ridge, US—Upper slope, MS—Middle slope, LS—Lower slope, VL—Valley, FS—Flat slope.

Technical Validation

After the integral version of the RLBJ inventory was prepared, a hybrid technical validation process was conducted to verify: (i) inventory compilation completeness and integrity; (ii) geographic accuracy, i.e., size, location, and shape of the individual landslides; and (iii) thematic accuracy, i.e., detection rate and landslide classification 13. Activities to validate inventory completeness and correctness were conducted for all landslide records of the dataset. Other validations were carried out based on field investigation on a random sampling of ~2% of the features originally mapped. Field investigations are mainly conducted along accessible roads in areas with a dense distribution of landslides. Due to the fact that many landslides occurred along high mountains, that were unreachable by roads, a telephoto lens with a maximum focal length of 200 millimeters was used to obtain high-definition photos of landslides from a distance. A total of 403 landslides were confirmed during field investigations (field photos and necessary validation logs are included in the “Field logs” document in the repository).

Inventory completeness and integrity were checked on the GIS platform. The attribute tables of all mapped features (i.e., points and polygons) were systematically checked to prevent any empty records, typing errors, exception values, and referential integrity issues. During the first checking process, empty records and wrong attribute values (i.e., polygon areas with negative values) of the key geometric properties were found in the attribute table. After checking for elements with incorrect attributes, we found that these errors were caused by (i) the wrong line crossing at polygon edges, (2) the topological overlap of adjacent polygons, and (3) the meaningless geometric elements left behind by errors during the vectorization process. Referential integrity issues existed regarding the point and polygon layers, indicating that the number of points and polygons was not equal in the raw version of the merged inventory. This is because the points and polygons are labeled separately. Considering the huge number of landslides, the entire study area was divided into multiple regions during the interpretation process; therefore, incorrect repetitive marking led to overlaps in the boundary zones of these regions during data merging. The topology-checking tool in ArcGIS can efficiently identify the mismatch between points and polygons. These errors were all identified and corrected one by one, and the final version of the inventory showed no empty records, wrong attributes, or referential integrity issues.

Geographic accuracy mainly involves the accuracy of landslide location and area limits. The geometric features of each identified landslide were marked based on the best visible satellite images, and at least two satellite images from other dates (before and after the landslide, respectively) were used to further confirm the location and boundary of the landslides. Therefore, we confirmed that the positions of the initial points of all identified landslides were accurate. However, the accuracy of the landslide boundaries requires additional validation, as it is controlled by the inherent limitations of the manual visual interpretation method (e.g., image resolution, and visual error). Image resolution and visual errors have a significant impact on landslide delineation, especially for landslides with small areas. After conducting the field investigation, we carefully checked the polygons of the landslides that could be identified in field photos. The area of the initial polygon was generally 5% to 10% larger than that of the verified polygon; for some small landslides, the maximum error can reach ~30%. In addition, the coordinate projection in the GIS platform may cause area measurement errors. We selected five landslides with clear boundaries and measured their geometric dimensions (i.e., landslide length and width) in the field using a handheld rangefinder. The deviation between the values measured in the field and those measured by GIS tools was less than 1 m.

For thematic accuracy, our study mainly focuses on the detection rate and classification accuracy of remote-sensing landslide identification. Therefore, our verification mainly relies on on-site data. We used the field photos, locations of the landslides depicted in the photos, and views of the satellite images to match the landslides identified through field investigations with those identified through remote sensing interpretation. If a landslide confirmed by field investigation was identified by remote sensing interpretation and the classification is consistent, then the landslide was deemed to be correctly identified (Table 2). For the 403 landslides identified in the field (Fig. 8a), 337 landslides (83.6%) were identified by remote sensing interpretation, among which 312 landslides were correctly classified (accounting for 77.4%), 24 landslides were misclassified (accounting for 6%), and one polygon was confirmed not to be a landslide but the artificially deposited gravels manually moved from the adjacent road surface during emergency efforts. However, 16.4% of landslides were not detected by remote sensing interpretation. The main reason for this omission is that the area of these landslides was too small, and the characteristics presented on satellite images cannot be used to fully determine whether these events could be considered landslides. Figure 8b further summarizes the thematic accuracy of each category of landslides. Landslides without flowing movements had the highest missed detection rate. This type of landslide mostly forms on steep road-cutting slopes, and the area it covers on satellite images is relatively small; therefore, only the landslides of this type with clear, recognizable features (similar to the cases shown in Table 2) were marked during interpreting. Similarly, the omission rate of landslides flowing on slopes with the same smaller area is also higher (18%, Fig. 8b). The proportion of misclassifications of different types of landslides is similar (Fig.

Results of the technical validation based on the field investigation. Note: “Correct” refers to the landslide identified in the field identified through remote sensing interpretation with a classification deemed correct, “Misclassified” refers to the landslides identified in the satellite images but with inaccurate classifications, “Misidentified” refers to cases where a feature is identified as a landslide in remote sensing images, but field investigations confirm it is not a landslide, and “Unidentified” refers to landslides not identified in the satellite images. The results in Panel b are disaggregated based on landslide type: 1—Slide-flow (Un-channeled), 2—Slide-flow (Channeled), 3—Slide-flow (Flood), 4—Slide/fall (Un-flow).

Although remote sensing interpretation has identified a large number of landslides, this inventory does not include all landslides triggered by this event. Based on the detection rate obtained from actual investigations, it can be estimated that the total landslide number in the study area may be nearly 20,000. We maintain this inventory to provide as complete a snapshot as possible of the massive landslides formed on both natural and artificial slopes triggered by the extreme precipitation that hit mountainous western regions of Beijing, China, from July 29, 2023 to August 2, 2023.

Usage Notes

All researchers interested in the disaster effects of climate change, rainfall-induced landslides, regional susceptibility and risk of landslides, and the application of machine learning methods in landslide analysis can benefit from this landslide inventory. This dataset can be used for any purpose for free. The final integration of this inventory was completed in ArcGIS 10.8. The dataset provided in the repository can be immediately reproduced in ArcGIS or other GIS platforms for further statistical analysis and integration with other databases. Despite the high number of features mapped, the current version of the RLBJ inventory still lacks many small-scale landslides and has limitations in landslide classification. Further versions of the RLBJ will strive to integrate more detailed metrics when higher-resolution satellite images or additional field validation data are available. The authors will continue to monitor rainfall-induced landslides in the same area in the future to evaluate the impact of long-term climate change on geological disasters in Northern China. Researchers are welcome to contribute to the integrity and richness of the mapped features or attributes in the inventory.

Code availability

No customized code was produced to prepare or analyze the dataset.

1.
Zhang, S. et al. Fatal landslides in China from 1940 to 2020: occurrences and vulnerabilities. Landslides 20, 1243–1264,
https://doi.org/10.1007/s10346-023-02034-6(2023).
2.
Froude, M. J. & Petley, D. N. Global fatal landslide occurrence from 2004 to 2016. Natural Hazards and Earth System Sciences 18, 2161–2181 (2018).
3.
Garcia-Delgado, H., Petley, D. N., Bermúdez, M. A. & Sepúlveda, S. A. Fatal landslides in Colombia (from historical times to 2020) and their socio-economic impacts. Landslides 19, 1689–1716,
https://doi.org/10.1007/s10346-022-01870-2(2022).
4.
Dille, A. et al. Acceleration of a large deep-seated tropical landslide due to urbanization feedbacks. Nat. Geosci. 15, 1048–1055,
https://doi.org/10.1038/s41561-022-01073-3(2022).
5.
Lavé, J. et al. Medieval demise of a Himalayan giant summit induced by mega-landslide. Nature 619, 94–101,
https://doi.org/10.1038/s41586-023-06040-5(2023).
6.
Pittau, S., Rossi, M., Llena, M. & Brardinoni, F. Evaluating historical, basin-wide landslide activity in a context of land abandonment and climate change: Effects of landslide visibility and temporal resolution. Geomorphology 452, 109122,
https://doi.org/10.1016/j.geomorph.2024.109122(2024).
7.
Mugaruka Bibentyo, T. et al. Landslides, bedrock incision and human-induced environmental changes in an extremely rapidly formed tropical river gorge. Geomorphology 449, 109046,
https://doi.org/10.1016/j.geomorph.2023.109046(2024).
8.
Losen, J. et al. Repeated failures of the giant Beshkiol Landslide and their impact on the long-term Naryn Basin floodings, Kyrgyz Tien Shan. Geomorphology 453, 109121,
https://doi.org/10.1016/j.geomorph.2024.109121(2024).
9.
Xie, J., Coulthard, T. J., Cao, J., Wu, J. H. & Wang, M. How does landslide debris grain size control sediment transport and dynamics? Geomorphology 451, 109050,
https://doi.org/10.1016/j.geomorph.2023.109050(2024).
10.
Kim, H., Lee, J. H., Park, H. J. & Heo, J. H. Assessment of temporal probability for rainfall-induced landslides based on nonstationary extreme value analysis. Engineering Geology 294, 106372,
https://doi.org/10.1016/j.enggeo.2021.106372(2021).
11.
Tsai, F., Hwang, J. H., Chen, L. C. & Lin, T. H. Post-disaster assessment of landslides in southern Taiwan after 2009 Typhoon Morakot using remote sensing and spatial analysis. Natural Hazards and Earth System Sciences 10, 2179–2190 (2010).
12.
Cabral, V. et al. The consequences of debris flows in Brazil: a historical analysis based on recorded events in the last 100 years. Landslides 20, 511–529,
https://doi.org/10.1007/s10346-022-01984-7(2023).
13.
Santangelo, M. et al. Inventory of landslides triggered by an extreme rainfall event in Marche-Umbria, Italy, on 15 September 2022. Sci Data 10, 427,
https://doi.org/10.1038/s41597-023-02336-3(2023).
14.
Westra, S., Alexander, L. V. & Zwiers, F. W. Global increasing trends in annual maximum daily precipitation. J. Clim. 26, 3904–3918,
https://doi.org/10.1175/JCLI-D-12-00502.1(2013).
15.
Fischer, E. M. & Knutti, R. Anthropogenic contribution to global occurrence of heavy-precipitation and high-temperature extremes. Nat. Clim. Chang. 5, 560–564,
https://doi.org/10.1038/nclimate2617(2015).
16.
Myhre, G. et al. Frequency of extreme precipitation increases extensively with event rareness under global warming. Sci. Rep. 9, 1–10,
https://doi.org/10.1038/s41598-019-52277-4(2019).
17.
Donat, M. G., Lowry, A. L., Alexander, L. V., O’Gorman, P. A. & Maher, N. More extreme precipitation in the world’s dry and wet regions. Nat. Clim. Chang. 6, 508–513,
https://doi.org/10.1038/nclimate2941(2016).
18.
Tian, S. F. et al. New insights into the occurrence of the catastrophic Zhaiban slope debris flow that occurred in a dry valley in the Hengduan Mountains in southwest China. Landslides 19, 647–657,
https://doi.org/10.1007/s10346-021-01824-0(2022).
19.
Wang, H. J. et al. Rainfall-induced landslide in loess area, Northwest China: a case study of the Changhe landslide on September 14, 2019, in Gansu Province. Landslides 17, 2145–2160,
https://doi.org/10.1007/s10346-020-01460-0(2020).
20.
Li, M. Y. et al. Landslide response to vegetation by example of July 25–26, 2013, extreme rainstorm, Tianshui, Gansu Province, China. Bulletin of Engineering Geology and Environment 80, 751–764,
https://doi.org/10.1007/s10064-020-02000-9(2021).
21.
Rüther, D. C., Lindsay, E. & Slåtten, M. S. Landslide inventory: ‘Hans’ storm southern Norway, August 7–9, 2023. Landslides 21, 1155–1159,
https://doi.org/10.1007/s10346-024-02222-y(2024).
22.
Reyes-Carmona, C. et al. Improving landslide inventories by combining satellite interferometry and landscape analysis: the case of Sierra Nevada (Southern Spain). Landslides 20, 1815–1835,
https://doi.org/10.1007/s10346-023-02071-1(2023).
23.
Duman, T. Y. et al. Landslide inventory of northwestern Anatolia, Turkey. Engineering Geology 77(1–2), 99–114 (2005).
24.
Huang, D., Ma, H. & Huang, R. Q. Deep-seated toppling deformations of rock slopes in western China. Landslides 19, 809–827 (2022).
25.
Fusco, F. et al. A revised landslide inventory of the Campania region (Italy). Sci Data 10, 355,
https://doi.org/10.1038/s41597-023-02155-6(2023).
26.
Sharma, N., Saharia, M. & Ramana, G. V. High resolution landslide susceptibility mapping using ensemble machine learning and geospatial big data. Catena 235, 107653,
https://doi.org/10.1016/j.catena.2023.107653(2024).
27.
Chen, X. & Chen, W. GIS-based landslide susceptibility assessment using optimized hybrid machine learning methods. Catena 196, 104833,
https://doi.org/10.1016/j.catena.2020.104833(2021).
28.
Cardinali, M. et al. Rainfall induced landslides in December 2004 in south-western Umbria, central Italy: types, extent, damage and risk assessment. Natural Hazards and Earth System Sciences 6, 237–260 (2006).
29.
Guzzetti, F., Reichenbach, P., Cardinali, M., Galli, M. & Ardizzone, F. Probabilistic landslide hazard assessment at the basin scale. Geomorphology 72, 272–299 (2005).
30.
Nguyen, B. Q. V. & Kim, Y. T. Regional-scale landslide risk assessment on Mt. Umyeon using risk index estimation. Landslides 18, 2547–2564,
https://doi.org/10.1007/s10346-021-01622-8(2021).
31.
Hovius, N., Stark, C. P. & Allen, P. A. Sediment flux from a mountain belt derived by landslide mapping. Geology 25, 231–234 (1997).
32.
Hovius, N., Stark, C. P., Chu, H. T. & Lin, J. C. Supply and removal of sediment in a landslide-dominated mountain belt: Central Range, Taiwan. Journal of Geology 108, 73–89 (2000).
33.
Guzzetti, F. et al. Distribution of landslides in the Upper Tiber River basin, central Italy. Geomorphology 96, 105–122 (2008).
34.
Parker, R. N. et al. Mass wasting triggered by the 2008 Wenchuan earthquake is greater than orogenic growth. Nature Geoscience 4(7), 449–452 (2011).
35.
Yang, F. et al. Modelling the evolution of debris flows after the 2008 Wenchuan earthquake. Engineering Geology 321, 107152,
https://doi.org/10.1016/j.enggeo.2023.107152(2023).
36.
Wang, X. et al. Long‐term landslide evolution and restoration after the Wenchuan earthquake revealed by time‐series remote sensing images. Geophysical Research Letters 51, e2023GL106422,
https://doi.org/10.1029/2023GL106422(2024).
37.
Gamboa, D., Omira, R. & Terrinha, P. A database of submarine landslides offshore West and Southwest Iberia. Sci Data 8, 185,
https://doi.org/10.1038/s41597-021-00969-w(2021).
38.
Guzzetti, F. et al. Landslide inventory maps: new tools for an old problem. Earth-Science Reviews 112, 42–66 (2012).
39.
Huang, Y. D. et al. An open-accessed inventory of landslides triggered by the MS 6.8 Luding earthquake, China on September 5, 2022. Earthquake Research Advances 3, 100181 (2023).
40.
Guzzetti, F. et al. Landslides triggered by the 23 November 2000 rainfall event in the Imperia Province, Western Liguria, Italy. Engineering Geology 73(2), 229–245 (2004).
41.
Cardinali, M., Ardizzone, F., Galli, M., Guzzetti, F. & Reichenbach, P. Landslides triggered by rapid snow melting: the December 1996–January 1997 event in Central Italy. In: Claps, P., Siccardi, F. (Eds.), Proceedings 1st Plinius Conference, Maratea. Bios Publisher, Cosenza, 439–448 (2000).
42.
Dai, F. C. et al. Spatial distribution of landslides triggered by the 2008 Ms 8.0 Wenchuan earthquake. Journal of Asian Earth Sciences 40(3), 883–895 (2010).
43.
Santangelo, M., Cardinali, M., Rossi, M., Mondini, A. C. & Guzzetti, F. Remote landslide mapping using a laser rangefinder binocular and GPS. Natural Hazards and Earth System Sciences 10, 2539–2546 (2010).
44.
Weirich, F. & Blesius, L. Comparison of satellite and air photo based landslide susceptibility maps. Geomorphology 87(4), 352–364 (2007).
45.
Crosta, G. B., Frattini, P. & Agliardi, F. Deep seated gravitational slope deformations in the European Alps. Tectonophysics 605, 13–33 (2013).
46.
Blondeau, S., Gunnell, Y. & Jarman, D. Rock slope failure in the Western Alps: A first comprehensive inventory and spatial analysis. Geomorphology 380, 107622,
https://doi.org/10.1016/j.geomorph.2021.107622(2021).
47.
Santurri, L. et al. Assessment of very high-resolution satellite data fusion techniques for landslide recognition. ISPRS TC VII Symposium. IAPRS, Vienna, Austria, XXXVIII (7B), 492–496 (2010).
48.
Cascini, L., Fornaro, G. & Peduto, D. Advanced low- and full-resolution DInSAR map generation for slow-moving landslide analysis at different scales. Engineering Geology 112, 29–42 (2010).
49.
Fiorucci, F. et al. Seasonal landslides mapping and estimation of landslide mobilization rates using aerial and satellite images. Geomorphology 129(1–2), 59–70 (2011).
50.
Martha, T. R., van Westen, C. J., Kerle, N., Jetten, V. & Kumar, K. V. Landslide hazard and risk assessment using semi-automatically created landslide inventories. Geomorphology 184, 139–150,
https://doi.org/10.1016/j.geomorph.2012.12.001(2013).
51.
Borghuis, A. M., Chang, K. & Lee, H. Y. Comparison between automated and manual mapping of typhoon-triggered landslides from SPOT-5 imagery. International Journal of Remote Sensing 28, 1843–1856 (2007).
52.
State Council Information Office of the People’s Republic of China. Press conference on flood prevention and disaster relief work in Beijing. Accessed 9 August 2023,
http://www.scio.gov.cn/xwfb/dfxwfb/gssfbh/bj_13826/202308/t20230811_750….
53.
JAXA/METI. ALOS PALSAR Radiometric_Terrain_Corrected_high_res. Accessed through ASF DAAC 20 March 2024.
https://doi.org/10.5067/Z97HFCNKR6VA(2015).
54.
Hungr, O., Leroueil, S. & Picarelli, L. The Varnes classification of landslide types, an update. Landslides 11(2), 167–194,
https://doi.org/10.1007/s10346-013-0436-y(2014).
55.
Weiss, A. Topographic position and landforms analysis. ERSI User Conference, San Diego, CA, USA (2001).
56.
Ma, H. & Wang, F. W. Inventory of shallow landslides triggered by extreme precipitation in July 2023 in Beijing, China. Figshare
https://doi.org/10.6084/m9.figshare.26878327(2024).

Acknowledgements

This research was funded by the National Natural Science Foundation of China (Grant No. 42230715) and the Fundamental Research Fund for the Central Universities of China.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

About this article

Cite this article

Ma, H., Wang, F. Inventory of shallow landslides triggered by extreme precipitation in July 2023 in Beijing, China. Sci Data 11, 1083 (2024). https://doi.org/10.1038/s41597-024-03901-0

-
Received:
-
Accepted:
-
Published:
-
DOI: https://doi.org/10.1038/s41597-024-03901-0