Broadly speaking, my area of study is the modeling of the natural world. Many models are vast simplifications of the complexity of nature. My objective is to develop sophisticated models that generalize less and explain more. To date, this has led in two directions:
A “sense of place” is sometimes palpable.
1. Collect, quantify, and visualize people’s perceptions of place. In this research, which I performed to complete my Ph.D. (2017), I examined how study participants rated features in a public park, what emotions they associated with them, and how far these feelings extended from the features in question. This allowed the perceptions to be mapped and analyzed, so that we could map place importance for individuals or groups.
2. Analyze digital elevation models to quantify their form and classify their component landforms. After early successes using a Fuzzy ArtMAP neural network to identify eskers (using BC TRIM data built into a TIN), I am now working with a team to go further and identify esker components with a convolutional neural network and corrected raster data from the ArcticDEM. In the decade and a half since I first approached this problem, the million-fold increase in the speed of computers, the availability of high resolution data over a huge expanses of the Arctic, and the advances in Neural Networks have made widespread use of this technology possible.