School of GeoSciences

School of GeoSciences


Sarah Touati


About Me

I have a BSc (first class honours) in Physics with Music from the University of Edinburgh, an MSc in Information Technology from Heriot-Watt University, and a PhD in statistical seismology from the University of Edinburgh. I have 2 years' experience in software engineering and have been working in a postdoctoral research position since October 2011.


I am currently working in the RACER consortium of the PURE project. My role here is to produce unified catalogue of UK earthquakes with moment magnitude estimates for all events, and full uncertainties on magnitude and location, for the purpose of an updated seismic hazard analysis. This involves the conversion of early instrumental and pre-instrumental information into moment magnitude. I am using Bayesian inference to capture the information and its uncertainty estimates (which are themselves uncertain!) and infer posterior distributions for the unknown quantities, according to a carefully constructed conceptual map of how the pieces of information relate to one another.

I am also involved in the REAKT project, for which I am developing techniques for using statistical models of earthquake occurrence (such as ETAS) to identify the existence or non-existence of changes in the basic rate of events within an earthquake time series. (This has relevance to, for example, the current question of whether the recent global cluster of megaquakes is statistically significant.) One of the methods I'm implementing for doing this involves the estimation of Bayesian evidence, or the Bayes Factor, which can be used to choose between models -- for example, between a stationary and a non-stationary model.

My PhD research explained the sometimes bimodal shape of the earthquake inter-event time distribution, in terms of the relative proportions of intervals that are between events related to each other (through aftershock triggering) and unrelated events. I went on to explore the effect of these relative proportions on the accuracy of parameter inversions for the ETAS model, based on 'synthetic' data from simulations of ETAS.


  1. Zhuang, J. C. and S. Touati (2014), Stochastic simulation of earthquake catalogues, Community Online Resource for Statistical Seismicity Analysis (in review, doi to come - will be available at
  2. Touati, S., Naylor, M., and I.G. Main (2014). Statistical modeling of the 1997-98 Colfiorito earthquake sequence: locating a stationary solution within parameter uncertainty, Bull. Seismol. Soc. Am., 104 (2), 885-897. (Link to article)
  3. Main, I.G., Bell, A.F., Meredith, P.G., Geiger, S., and S. Touati (2012). The dilatancy-diffusion hypothesis and earthquake predictability, in Healy, D., Butler, R. W. H., Shipton, Z. K. & Sibson, R. H. (eds). Faulting, Fracturing and Igneous Intrusion in the Earth’s Crust. Geological Society, London, Special Publications, 367, 215–230. (Link to article)
  4. Main, I.G., Naylor, M., Greenhough, J., Touati, S., Bell, A.F. & J. McCloskey (2011). Model selection and uncertainty in earthquake hazard analysis, in Köhler & Nishijima (eds), Applications of Statistics and Probability in Civil Engineering, Faber, Taylor & Francis Group, London, ISBN 978-0-415-66986-3, pp. 735-743. (Link to article)
  5. Touati, S., Naylor, M., Main, I.G., and M. Christie (2011). Masking of earthquake triggering behavior by a high background rate and implications for ETAS inversions, J. Geophys. Res., 116, B03304. (Link to article)
  6. Touati, S., Naylor, M., and I.G. Main (2009). Origin and nonuniversality of the earthquake interevent time distribution, Phys. Rev. Lett. 102, 168501. (Link to article)
  7. Naylor, M., Main, I.G., and S. Touati (2008). Quantifying uncertainty in mean earthquake interevent times for a finite sample, J. Geophys. Res., 114, B01316. (Link to article)

Recent Posters

R codes

PhD Thesis

Complexity, aftershock sequences, and uncertainty in earthquake statistics

Contact Details

Available here