PhD Public Lecture (DSAS) - Chelsea Uggenti

This thesis considers and articulates principles of interdisciplinary knowledge exchange in two contexts: the development of novel techniques for the study of wildland fire lifetimes; and, to understand improvements in the student training environment focusing on graduate teaching assistants, developing a training program on active learning for graduate teaching assistants in the mathematical and statistical sciences.

Wildland fire science is an area of research that requires interdisciplinary expertise to advance its body of knowledge. Wildland fires that are suppressed have a "lifetime" that consists of several sequential phases, including what are called detection and action phases. The interconnectedness of these phases is often overlooked when studying fire responses, and we develop methods to fill that gap in this thesis. In particular, we consider such a framework for the analysis of fire data from the Sioux Lookout District in northwestern Ontario. Multi-state modelling and joint frailty modelling techniques are employed. Comparisons of different frailty distributions and random effect forms are considered, and a simulation study is performed to highlight the advantages of a flexible model form for the joint frailty models. Using the joint frailty models, we find that fires with longer detection phases are associated with longer action phases, and that the action phase lengths may be increasing over time. Collaboration with fire scientists throughout the development of this work was critical and is especially important for ensuring the impact of it at fire management agencies.

The importance of collaboration in statistics is emphasized in how education in this field is conducted. A workshop on active learning techniques, which aid in the exchange of knowledge between students and instructors, was developed for graduate teaching assistants at the University of Western Ontario. A mixed-method study was performed to answer the question of the importance and impact of active learning techniques. Learnings from this study are discussed.