Geostatistical Modeling to Capture Seismic‐Shaking Patterns From Earthquake‐Induced Landslides. Issue 7 (24th July 2019)

• Record Type: Journal Article
• Authors: Lombardo, Luigi; Bakka, Haakon; Tanyas, Hakan; van Westen, Cees; Mai, P. Martin; Huser, Raphaël
• Is Part Of: Journal of geophysical research. Volume 124:Issue 7(2019)
• Page Start: 1958
• Page End: 1980
• Publication Date: 2019-07-24
• Subjects: Integrated Nested Laplace Approximation (INLA) -- landslide susceptibility -- landslide intensity -- slope unit -- spatial point pattern -- Wenchuan and Lushan earthquakes; Geomorphology -- Periodicals; 551.3
• DOI: 10.1029/2019JF005056
• Languages: English
• ISSNs: 2169-9003

Abstract

Abstract : We investigate earthquake‐induced landslides using a geostatistical model featuring a latent spatial effect (LSE). The LSE represents the spatially structured residuals in the data, which remain after adjusting for covariate effects. To determine whether the LSE captures the residual signal from a given trigger, we test the LSE in reproducing the pattern of seismic shaking from the distribution of seismically induced landslides, without prior knowledge of the earthquake being included in the model. We assessed the landslide intensity, that is, the expected number of landslides per mapping unit, for the area in which landslides triggered by the Wenchuan and Lushan earthquakes overlap. We examined this area to test our method on landslide inventories located in near and far fields of the earthquake. We generated three models for both earthquakes: (i) seismic parameters only (proxy for the trigger); (ii) the LSE only; and (iii) both seismic parameters and the LSE. The three configurations share the same morphometric covariates. This allowed us to study the LSE pattern and assess whether it approximated the seismic effects. Our results show that the LSE reproduced the shaking patterns for both earthquakes. In addition, the models including the LSE perform better than conventional models featuring seismic parameters only. Due to computational limitations we carried out a detailed analysis for a relatively small area (2, 112 km 2 ), using a data set with higher spatial resolution. Results were consistent with those of a subsequent analysis for a larger area (14, 648 km 2 ) using coarser‐resolution data. Key Points: Landslide inventories are used to estimate the ground motion patterns via spatial statistics without any prior knowledge of the earthquake A spatial point process jointly predicts the location and number of landslides; we defined it as landslide intensity The Poisson aggregative property produces landslide intensity maps for any mapping unit y using a single model