ETAS.inlabru

R package that implements the ETAS Hawkes process for modelling seismicity

Online documentation: https://edinburgh-seismicity-hub.github.io/ETAS.inlabru/

Authors

• Dr Francesco Serafini
• Dr Mark Naylor , School of GeoSciences, University of Edinburgh
• Prof Finn Lindgren , School of Mathematics, University of Edinburgh
• Dr Kirsty Bayliss , Global Earthquake Model (GEM)

Funding

• This study was funded by yhe Real-Time Earthquake Risk Reduction for a Resilient Europe RISE project , which has received funding from the European Union’s Horizon 2020 Research and Innovation Program under grant Agreement 821115.
• Naylor was additionally funded by the NSFGEO-NERC grant NE/R000794/1.
• Bayliss was funded by an EPSRC Studentship.

Installation

For ETAS.inlabru to work, we need to install both R-INLA and inlabru:

For inlabru (see https://inlabru-org.github.io/inlabru/):

• CRAN release,

install.packages("inlabru")

• or development version,

# install.packages("remotes")
remotes::install_github("inlabru-org/inlabru")

For R-INLA (see https://www.r-inla.org/download-install):

install.packages(
  "INLA",
  repos = c(getOption("repos"), INLA="https://inla.r-inla-download.org/R/testing"),
  dep = TRUE
)

You can install the development version of ETAS.inlabru from GitHub with

# install.packages("remotes")
remotes::install_github("edinburgh-seismicity-hub/ETAS.inlabru")

Terminology and planning suggestions

File structure in package

• ETAS.triggering.function.R : Contains the ETAS specific model functions
• HawkesProcess.R : Generic Hawkes code that is intended for integration back into inlabru
• generateSyntheticCatalogues.R : Contains the iterative Hawkes functions for generating triggered events but the actual triggering functions reside in the ETAS file above so we could introduce other models here
• temporalBinning.R : Code to generate time bins to make integration scheme efficient
• plottingFunctions.R : lets put all the standard plotting functions into here
• setupInlabruInputs.R : Put the functions for generating input.list in here

Terminology

• Let’s be specific about when we are doing just temporal so we have clear function names for the spatial and spatial-temporal later on
• I changed the theta from c(mu, K, ...) to a df <- data.frame(mu=mu, K=K, alpha=alpha, c=c, p=p). This means we should refer to the values as df$mu which is unambiguous etc

What has been done

Implemented

• Generation of synthetic ETAS catalogues by ETAS.inlabru with a demonstration in the notebook

In development

• Add inversion modelling based on original code
• Modify the implementation so that the generic Hawkes code can go into inlabru and the ETAS triggering function code stay in this package

Roadmap

• Integrate spatial modelling