An exploration of space-time models for landslide hazard and prediction

Erin Bryce (University of Glasgow)

Wednesday 7th May 14:00-15:00 Maths 311B

Abstract

This work follows an exploration of spatial and spatio-temporal statistical models within a Bayesian framework, applied to large datasets of a specific geomorphological discretisation. We show how complex geostatistical, hurdle, and point process models can be fitted to irregular resolutions while efficiently leveraging available environmental data. Additionally, we illustrate how statistical results can be translated into risk measures for effective communication with stakeholders. Most of the inference is conducted using the integrated nested Laplace approximation (INLA) with the stochastic partial differential equation (SPDE) approach, though frequentist methods are also explored in the context of functional data.

Add to your calendar

Download event information as iCalendar file (only this event)