Graham Shapiro

Utilizing long-term memory to model Northern California earthquakes

Does there exist a relationship between the times of past earthquakes and the time until the next earthquake? Recent research suggests that there exists patterns in earthquake occurrence that exist between long sequences of earthquakes. Most current earthquake models assume earthquakes are either memoryless or only incorporate short-term memory. The purpose of this research is to develop a statistical earthquake model that incorporates time-dependence as well as long-term memory. Specifically to create a computer implemented time dependant statistical model of temporal sequences of earthquake features to estimate the effect of long term memory on the magnitude and occurrence times of earthquakes in Northern California.

Message to Sponsor

I believe that investigating earthquage phenomena is a worthwhile endeavor even if only a small amount of novel predictive information is uncovered. The possibility of modeling with greater accuracy the dynamics of earthquake activity has tremendous social and economic implications. Additionally, the potential of contributing to such an significant and relevant area of research is exciting as well as humbling. I will be building upon the work of brilliant scientists and statisticians as well as contributing to a new method of earth analysis.
  • Major: Statistics
  • Mentor: Cari Kaufman, Statistics