This repository documents and references the geospatial datasets used for the joint modelling of earthquake-induced landslides (EQIL), centroid locations, and landslide sizes in square metres in Nepal following the 2015 Gorkha earthquake. The data include administrative boundaries, geology, elevation, land cover, rainfall, and landslide inventories.
fit6a is the main landslide susceptibility model used in this study. It estimates how landslide susceptibility varies across the study area.
Landslide susceptibility ~ PGA effect + river-incision effect + rainfall effect + channel-distance effect
In simplified form, the model is:
log(lambda(s)) ~ beta0 + f1(PGA(s)) + f2(log(ksn(s))) + f3(Rainfall(s)) + beta * exp(-Fd2Ch(s))
The model combines information on:
- earthquake shaking;
- topography, represented by
log(ksn); - rainfall; and
- distance to the nearest fluvial channel.
Here, lambda(s) is the predicted landslide intensity or susceptibility at location s. The terms f1, f2, and f3 allow flexible nonlinear relationships with shaking, channel steepness, and rainfall. The channel-distance term gives a stronger effect close to fluvial channels and a weaker effect farther away. The spatial random effect u(s) captures remaining spatial clustering not explained by the other variables.
In simple terms, fit6a predicts where landslides are more likely to occur based on shaking, landscape form, rainfall, proximity to channels, and unresolved spatial structure.
If you use this dataset, code, or analysis, please cite:
Suen, M. H., Naylor, M., Mudd, S., & Lindgren, F. (2026). Influence of river incision on landslides triggered in Nepal by the Gorkha earthquake: Results from a pixel-based susceptibility model using inlabru. Accepted for publication in Frontiers in Earth Science: Geohazards and Georisks.
Additional technical details are provided in the accepted manuscript.
For statistical details related to spatial misalignment, see:
Suen, M. H., Naylor, M., & Lindgren, F. (2026). Coherent disaggregation and uncertainty quantification for spatially misaligned data. Environmetrics, 37(2), e70078. https://onlinelibrary.wiley.com/doi/abs/10.1002/env.70078
@article{suen2025influence,
title={Influence of river incision on landslides triggered in Nepal by the Gorkha earthquake: Results from a pixel-based susceptibility model using inlabru},
author={Suen, Man Ho and Naylor, Mark and Mudd, Simon and Lindgren, Finn},
journal={arXiv preprint arXiv:2507.08742},
year={2025}
}
@article{suen2026coherent,
title={Coherent disaggregation and uncertainty quantification for spatially misaligned data},
author={Suen, Man Ho and Naylor, Mark and Lindgren, Finn},
journal={Environmetrics},
volume={37},
number={2},
pages={e70078},
year={2026},
publisher={Wiley Online Library}
}Note: Some datasets are updated periodically. Always verify the latest versions via official portals (e.g. USGS, FAO, etc.).
Main source:
FAO (2021):
The Himalaya Regional Land Cover Database
- Channel steepness index and distance metric to channel raster maps of the Gorkha Earthquake 2015-affected area computed from DEM and processed with LSDTopoTools (doi:10.5281/zenodo.8076231), see
lsdtopotools_driverfolder for scripts and details.
- To fill the gaps caused by buffering between the study area and
Nep_geology.shp, use the scriptnepal_geo_rast_fill.R. This applies nearest-neighbour interpolation to ensure full coverage in the Gorkha district, making the raster suitable for subsequent spatial analysis.
compiler.R: compiles and processes various geospatial datasets into a unified format for analysis and INLA spatial modelling viainlabru.tile_ldsize.R: Plots the landslide inventory with PGA contour lines and histogram for landslides.mchi.R: Plots the normalised channel steepness index (ksn) and channel profile analysis.pred_zm.R: Plots the posterior susceptibility map zoom-out.coefvar.R: Plots the coefficient of variation for the intensity and covariate effect maps.summary_stat.R: Summary statistics provided in Table 1.
The code is currently developed and tested in R 4.6.0. Below is the session information for reproducibility:
> sessionInfo()
R version 4.6.0 (2026-04-24)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 24.04.4 LTS
Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so; LAPACK version 3.12.0
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
time zone: Etc/UTC
tzcode source: system (glibc)
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] future_1.70.0 tidyterra_1.1.0 terra_1.9-27 ggplot2_4.0.3
[5] here_1.0.2 inlabru_2.14.1 INLA_26.05.10 Matrix_1.7-5
[9] fmesher_0.7.0 stars_0.7-2 sf_1.1-1 abind_1.4-8
[13] dplyr_1.2.1 patchwork_1.3.2
loaded via a namespace (and not attached):
[1] gtable_0.3.6 lattice_0.22-9 yyjsonr_0.1.22 vctrs_0.7.3
[5] tools_4.6.0 generics_0.1.4 curl_7.1.0 parallel_4.6.0
[9] tibble_3.3.1 proxy_0.4-29 pkgconfig_2.0.3 KernSmooth_2.23-26
[13] data.table_1.18.4 RColorBrewer_1.1-3 S7_0.2.2 lifecycle_1.0.5
[17] compiler_4.6.0 farver_2.1.2 stringr_1.6.0 textshaping_1.0.5
[21] codetools_0.2-20 nanoarrow_0.8.0 class_7.3-23 pillar_1.11.1
[25] tidyr_1.3.2 classInt_0.4-11 wk_0.9.5 parallelly_1.47.0
[29] gdalraster_2.6.1 tidyselect_1.2.1 digest_0.6.39 stringi_1.8.7
[33] purrr_1.2.2 listenv_0.10.1 labeling_0.4.3 splines_4.6.0
[37] rprojroot_2.1.1 grid_4.6.0 cli_3.6.6 magrittr_2.0.5
[41] maptiles_0.11.0 e1071_1.7-17 withr_3.0.2 scales_1.4.0
[45] sp_2.2-1 bit64_4.8.0 globals_0.19.1 bit_4.6.0
[49] otel_0.2.0 ragg_1.5.2 ggspatial_1.1.10 splancs_2.01-45
[53] viridisLite_0.4.3 rlang_1.2.0 isoband_0.3.0 Rcpp_1.1.1-1.1
[57] glue_1.8.1 DBI_1.3.0 xml2_1.5.2 R6_2.6.1
[61] systemfonts_1.3.2 units_1.0-1
The authors gratefully acknowledge funding from the EPSRC–UKRI Mathematics Doctoral Training Partnership
(grant number EP/W523847/1, project reference 2617239).