Please use this identifier to cite or link to this item:
|Title:||Calibration of Logistic Regression Coefficients from Limited Landslide Inventory Data for European Wide Landslide Susceptibility Modelling|
|Authors:||VAN DEN EECKHAUT MIET; HERVAS DE DIEGO Francisco; JAEDICKE Christian; MALET Jean-Philippe; PICARELLI Luciano|
|Citation:||Proceedings of the International conference 'Mountain Risks: Bringing Science to Society' p. 515-521|
|Type:||Articles in periodicals and books|
|Abstract:||In the collaborative activity of the EU-FP7 SafeLand project dealing with the identification of landslide hazard and risk hotspots in Europe, we applied logistic regression for landslide susceptibility modelling. The main focus was put on the calibration of the regression coefficients from a limited landslide inventory. First a representative sample of landslide-affected and landslide-free grid cells was created. Given that currently no European landslide inventory is available, landslide locations were collected from analysis of Google Earth images, and extracted from scientific publications and landslide databases available for Norway, a region around Naples, Italy and the Barcelonnette Basin, France. Then, a procedure taking account of the incompleteness of the landslide inventory and the high proportion of plain areas in Europe was set up to select landslide-free grid cells. Topographical, lithological and soil parameters were selected as input variables for the stepwise logistic regression. The model with the highest area under the ROC curve was selected for producing the landslide susceptibility map of Europe. The map was reclassified and validated by visual comparison with national landslide inventory or susceptibility maps available in literature. First results are promising, although low gradient areas known to be affected by landslides are not always classified as highly susceptible. Further validation is needed to fully evaluate and improve the methodology followed and the landslide susceptibility map produced.|
|JRC Institute:||Sustainable Resources|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.