Citation:

PANGAEA.935704

Name: Giant landslide inventory of Patagonia classified with the convolutional neural network AlexNet
Published: 2021-09-13
License: https://creativecommons.org/licenses/by/4.0/
Keywords: Convolutional neural network; Landslide detection; Landslide inventory; Landslides; Patagonia

Always quote citation above when using data! You can download the citation in several formats below.

Published: 2021-09-13 • DOI registered: 2022-07-12

RIS Citation BibTeX Citation Text Citation Copy Citation ShareShow Map Google Earth 90 29 14

Abstract:

We used the convolutional neural network AlexNet to detect giant landslides (>10^8 m³) along basaltic plateaus in the Patagonian extra-Andean region east of the Andean Cordillera (40°S-53°S, 66°W-72°W). The network was trained using topographic information (elevation, roughness, curvature) from TanDEM-X data. The dataset includes the original raster dataset as well as a polygon dataset. Since the network was trained with terrestrial data, large water bodies, the ocean as well as human settlements are sometimes detected as landslides. We removed the falsely predicted landslides patches in the polygon file of the dataset. Using artificial intelligence can help to analyze large quantities of data within a short time. The dataset shows are widespread landslides in the region are and how they might have been underestimated in their size and number in landslide inventories.

Keyword(s):

Convolutional neural network; Landslide detection; Landslide inventory; Landslides; Patagonia

Related to:

Schönfeldt, Elisabeth; Winocur, Diego; Pánek, Tomáš; Korup, Oliver (2022): Deep learning reveals one of Earth's largest landslide terrain in Patagonia. Earth and Planetary Science Letters, 593, 117642, https://doi.org/10.1016/j.epsl.2022.117642

Project(s):

Natural Hazards Group (NatHazGr)

StRATEGy international research training group (StRATEGy)

Funding:

Deutsche Forschungsgemeinschaft, Bonn (DFG), grant/award no. STR 373/34-1: StRATEGy

Coverage:

Latitude: -46.500000 * Longitude: -69.000000

Event(s):

Patagonia_region * Latitude: -46.500000 * Longitude: -69.000000

Comment:

The Landslide prediction raster dataset contains the original landslide prediction (pixel value = 1).

Landslide CNN prediction polygons cleaned, is the shapefile where contiguous landslide pixels are transformed to polygons, but removed depending on their location relative to water bodies, the ocean and human settlements.

Parameter(s):

#	Name	Short Name	Unit	Principal Investigator
1	File content	Content		Schönfeldt, Elisabeth
2	Binary Object (File Size)	Binary (Size)	Bytes	Schönfeldt, Elisabeth
3	Binary Object	Binary		Schönfeldt, Elisabeth

License:

Creative Commons Attribution 4.0 International (CC-BY-4.0)

Status:

Curation Level: Basic curation (CurationLevelB)

Size:

18 data points

Download Data

Download dataset as tab-delimited text — use the following character encoding:

View dataset as HTML

Datasets with similar metadata

Meier, WJ-H (2018): Glacier inventory for the Patagonian Andes, link to shape files. https://doi.org/10.1594/PANGAEA.891211
Franco, C; Maldonado, A; Ohlendorf, C et al. (2024): Bulk chemistry determination of Laguna Meseta (LME), Central Patagonia. https://doi.org/10.1594/PANGAEA.961935
Franco, C; Maldonado, A; Ohlendorf, C et al. (2024): Magnetic susceptibility of Laguna Meseta (LME), Central Patagonia. https://doi.org/10.1594/PANGAEA.961938