ArcGIS REST Services Directory | Login |
Home > services > puerto_rico_earthquake_202001 > coseismic_landslides (MapServer) | Help | API Reference |
JSON | SOAP | WMS |
Summary:
One hundred twenty landslides were mapped and inferred to have occurred during the period of strong ground shaking related to the January 7th, 2020 Puerto Rico Mw6.4 earthquake. Landslides are widely dispersed with the highest concentration in the southwestern portion of the island, nearest the epicenter. Landslides are widely distributed up to ~ 40 km north of the epicenter. Landslide size distribution is smaller than what is observed for other recent coseismic events. Most failures appear to be shallow debris slides, involving soil and weathered rock, without significant runout. Landslides primarily occur in Tertiary age limestone units.
Mapping methodology: Landslides were identified manually by an experienced geologist/GIS specialist using visual inspection of pre-/post-event high-resolution optical satellite imagery. Change in vegetation, surface roughness/surface morphology and albedo, and/or downslope appearance of loose debris were used to identify the perimeter of co-seismic landslides. Landslide polygon boundaries were manually drawn and registered to the post-event imagery. Boundaries include both source and deposit areas.
Various attribute fields are included in the polygon feature class, including: author, pre-event image filename, the post-event image filename, the date the landslide was mapped, and the confidence level (high/low). Landslides with low confidence had either low image quality in either the pre or post event imagery, or more rarely, were missing or had an uncertain date of pre-event imagery. A few landslides (n=10) were reactivations of previous landslides, but had clear evidence of new landslide activity observed between the pre and post image, such as new debris, surface roughness change or retrogression. The extent of landslide mapping is defined by a separate polygon feature class that defines the geographic limit of the imagery used and areas masked by clouds.
Limitations and uncertainties: Minimum landslide size is ~ 10 m2 based on imagery resolution. Landslides smaller than this value may not be represented by this dataset. This dataset is derived from imagery analysis only and has not been field-verified. Therefore, small landslides may be mistaken for other objects that are not well resolved by the resolution of imagery used or optical blurring/ misregistration.
Satellite/Sensor:
Orthorectified, multi-spectral images were used in GeoTiff format from WorldView-2 (WV03), WorldView-3 (WV03), and GeoEye-1 (GE01). These satellites provide 46cm, 31cm, and 41cm ground-sample distance resolution for pansharpened images respectively.
Resolution:
Landslide mapping was performed at 1:500 scale.
Credits:
Logan Knoper, Marin Clark, University of Michigan; Dimitrios Zekkos, University of California, Berkeley; Dalia Kirschbaum (NASA-Goddard), Preliminary map of January 7, 2020 Mw6,4 Puerto Rico co-seismic landslides. This work was supported by NASA’s A.37 Disasters Landslides Project.
Use Constraints:
Digital images used for mapping are restricted by Digital Globe Inc., however, the derived mapping products presented here are publicly available. US National Geospatial-Intelligence Agency (NGA) licenses DigitalGlobe imagery, which can be used for US federal government purposes per the NextView license agreement.